Data_preprocessing_and_E2G_model_training.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing necssary modules\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import os\n",
    "import pickle\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.preprocessing import StandardScaler \n",
    "from sklearn.dummy import DummyClassifier\n",
    "from sklearn.metrics import (\n",
    "        f1_score, classification_report, \n",
    "        confusion_matrix, roc_curve, \n",
    "        roc_auc_score, accuracy_score,\n",
    "        log_loss)\n",
    "from sklearn import __version__ as sklearn_version\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from imblearn.under_sampling import NearMiss\n",
    "from datetime import datetime\n",
    "from scripts.utils import (\n",
    "        load_device_data, \n",
    "        load_device_data_multi_label, \n",
    "        load_all_data,\n",
    "        load_all_data_class,\n",
    "        load_all_data_multi_class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Data folder path and Extention of the data files\n",
    "base_directory = '../rawdata'\n",
    "file_extension = \"*.csv\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Loading Device Data into pandas Dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Doorbells\n",
    "danmini_doorbell_df = load_device_data(base_directory, file_extension, 'Danmini_Doorbell')\n",
    "ennio_doorbell_df = load_device_data(base_directory, file_extension, 'Ennio_Doorbell')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Ecobee Thermostat\n",
    "ecobee_thermostat_df = load_device_data(base_directory, file_extension, 'Ecobee_Thermostat')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Web cam\n",
    "samsung_cam_df = load_device_data(base_directory, file_extension, 'Samsung_SNH_1011_N_Webcam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Baby monitor\n",
    "baby_monitor_df = load_device_data(base_directory, file_extension, 'Philips_B120N10_Baby_Monitor')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Security cam\n",
    "provision_cam1_df = load_device_data(base_directory, file_extension, 'Provision_PT_737E_Security_Camera')\n",
    "provision_cam2_df = load_device_data(base_directory, file_extension, 'Provision_PT_838_Security_Camera')\n",
    "simplehome_cam1_df = load_device_data(base_directory, file_extension, 'SimpleHome_XCS7_1002_WHT_Security_Camera')\n",
    "simplehome_cam2_df = load_device_data(base_directory, file_extension, 'SimpleHome_XCS7_1003_WHT_Security_Camera')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Checking out Missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_doorbell_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ecobee_thermostat_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ennio_doorbell_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "baby_monitor_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "provision_cam1_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "provision_cam2_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "samsung_cam_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "simplehome_cam1_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MI_dir_L5_weight         0\n",
       "MI_dir_L5_mean           0\n",
       "MI_dir_L5_variance       0\n",
       "MI_dir_L3_weight         0\n",
       "MI_dir_L3_mean           0\n",
       "                        ..\n",
       "HpHp_L0.01_radius        0\n",
       "HpHp_L0.01_covariance    0\n",
       "HpHp_L0.01_pcc           0\n",
       "label                    0\n",
       "device                   0\n",
       "Length: 117, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "simplehome_cam2_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Creating dict to store values of DF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataframe = {\"Danmin Door Bell\": danmini_doorbell_df, \n",
    "            \"Ecobee_Thermostat\": ecobee_thermostat_df,\n",
    "            \"Ennio_Doorbell\": ennio_doorbell_df,\n",
    "            \"Philips_B120N10_Baby_Monitor\": baby_monitor_df,\n",
    "            \"Provision_PT_737E_Security_Camera\": provision_cam1_df,\n",
    "            \"Provision_PT_838_Security_Camera\": provision_cam2_df,\n",
    "            \"Samsung_SNH_1011_N_Webcam\": samsung_cam_df,\n",
    "            \"SimpleHome_XCS7_1002_WHT_Security_Camera\": simplehome_cam1_df,\n",
    "            \"SimpleHome_XCS7_1003_WHT_Security_Camera\": simplehome_cam2_df\n",
    "            }"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Dummy classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dummy_classifier(data, scaling=False):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        \n",
    "    \n",
    "    strat = [\"most_frequent\", \"stratified\", \"constant\", \"uniform\"]\n",
    "    \n",
    "    for s in strat:\n",
    "        if s == 'constant':\n",
    "            dummy_clf = DummyClassifier(strategy =s , constant='benign')\n",
    "        else:\n",
    "            dummy_clf = DummyClassifier(strategy=s)\n",
    "        \n",
    "        dummy_clf.fit(X_train, y_train)\n",
    "        print(s)\n",
    "        print(dummy_clf.predict(X_test))\n",
    "        print(dummy_clf.score(X_test, y_test))\n",
    "        print(\"----------------------Done----------------------\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.6403744803430554\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'gafgyt' 'gafgyt']\n",
      "0.5097319061180399\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.049111263871157815\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['mirai' 'benign' 'mirai' ... 'gafgyt' 'benign' 'benign']\n",
      "0.33316638842515306\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.6138904064794248\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'gafgyt' 'mirai' ... 'mirai' 'mirai' 'gafgyt']\n",
      "0.5139155298030411\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.01601512184811954\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['benign' 'mirai' 'gafgyt' ... 'mirai' 'benign' 'gafgyt']\n",
      "0.3328640987705523\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.8883919362400375\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.8033380215658696\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.11160806375996249\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'gafgyt' 'benign' ... 'gafgyt' 'gafgyt' 'benign']\n",
      "0.49922175339896857\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5558973798861664\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'mirai' 'mirai' ... 'benign' 'mirai' 'gafgyt']\n",
      "0.4158839091758595\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.15927597966044102\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['benign' 'benign' 'benign' ... 'mirai' 'benign' 'gafgyt']\n",
      "0.33300263346318615\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5276684454961807\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['gafgyt' 'gafgyt' 'mirai' ... 'mirai' 'gafgyt' 'mirai']\n",
      "0.4426911034377289\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.07549159281707032\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['benign' 'gafgyt' 'gafgyt' ... 'gafgyt' 'gafgyt' 'mirai']\n",
      "0.3339732290182632\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5142949320502812\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['gafgyt' 'mirai' 'benign' ... 'mirai' 'gafgyt' 'mirai']\n",
      "0.4125814520368984\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.11828667930600474\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'mirai' 'benign' ... 'mirai' 'benign' 'mirai']\n",
      "0.3317666926888333\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.8606696456332673\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'gafgyt' 'gafgyt' 'benign']\n",
      "0.7610045572858831\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.1393303543667327\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['benign' 'gafgyt' 'benign' ... 'gafgyt' 'gafgyt' 'benign']\n",
      "0.49856529888866186\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5960558789110023\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['gafgyt' 'gafgyt' 'mirai' ... 'mirai' 'gafgyt' 'gafgyt']\n",
      "0.4812777840002781\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.05428380523488222\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'mirai' 'gafgyt' 'benign']\n",
      "0.3328286671018125\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.6065904532062935\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'gafgyt' 'gafgyt' ... 'mirai' 'gafgyt' 'mirai']\n",
      "0.5056572431517583\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.02313828120102802\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'benign' 'mirai' ... 'gafgyt' 'benign' 'gafgyt']\n",
      "0.33311132702313045\n",
      "----------------------xxxxxxx----------------------\n"
     ]
    }
   ],
   "source": [
    "# without scaling\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dummy_classifier(dataframe[k])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "| Device | Most Frequent | Stratified | Constant | Uniform |\n",
    "|---|---|---|---|---|\n",
    "| Danmin Door Bell  | 0.64 | 0.51 | 0.049 | 0.33 |\n",
    "| Ecobee_Thermostat | 0.61  | 0.51 | 0.016 | 0.33 |\n",
    "| Ennio_Doorbell | 0.89 | 0.80 | 0.11 | 0.50 |\n",
    "| Philips_B120N10_Baby_Monitor | 0.55 | 0.41 | 0.15 | 0.33 |\n",
    "| Provision_PT_737E_Security_Camera | 0.52 | 0.44 | 0.075  | 0.33 |\n",
    "| Provision_PT_838_Security_Camera | 0.51  | 0.41 | 0.11 | 0.33  |\n",
    "| Samsung_SNH_1011_N_Webcam | 0.86 | 0.76 | 0.13 | 0.49  |\n",
    "| SimpleHome_XCS7_1002_WHT_Security_Camera | 0.59 | 0.48 | 0.054 | 0.33 |\n",
    "| SimpleHome_XCS7_1003_WHT_Security_Camera | 0.60  | 0.50 | 0.023 | 0.33 |"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.6403744803430554\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'gafgyt' 'mirai' ... 'mirai' 'gafgyt' 'mirai']\n",
      "0.5086320337817932\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.049111263871157815\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'benign' 'benign' ... 'mirai' 'gafgyt' 'benign']\n",
      "0.33267210055975643\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.6138904064794248\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5133492580643875\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.01601512184811954\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['mirai' 'benign' 'benign' ... 'mirai' 'benign' 'mirai']\n",
      "0.33324693036851527\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.8883919362400375\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'gafgyt' 'benign' 'gafgyt']\n",
      "0.8046507266760431\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.11160806375996249\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['benign' 'benign' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.4984903891233005\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5558973798861664\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['benign' 'gafgyt' 'mirai' ... 'gafgyt' 'mirai' 'mirai']\n",
      "0.41616606594580163\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.15927597966044102\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'benign' 'benign' ... 'mirai' 'benign' 'benign']\n",
      "0.3327629519059235\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5276684454961807\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['gafgyt' 'benign' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.4416729046434694\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.07549159281707032\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['benign' 'benign' 'benign' ... 'mirai' 'benign' 'gafgyt']\n",
      "0.33350638688334583\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5142949320502812\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'gafgyt' 'mirai' ... 'mirai' 'gafgyt' 'gafgyt']\n",
      "0.41229069415457165\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.11828667930600474\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'gafgyt' 'mirai' ... 'benign' 'mirai' 'mirai']\n",
      "0.33324039702391384\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['gafgyt' 'gafgyt' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.8606696456332673\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['benign' 'gafgyt' 'gafgyt' ... 'benign' 'gafgyt' 'gafgyt']\n",
      "0.7606936313484414\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.1393303543667327\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'benign' 'gafgyt' ... 'gafgyt' 'gafgyt' 'gafgyt']\n",
      "0.49943589151349865\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.5960558789110023\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.4796672292665217\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.05428380523488222\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['benign' 'mirai' 'mirai' ... 'gafgyt' 'mirai' 'benign']\n",
      "0.3333693809212991\n",
      "----------------------xxxxxxx----------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "most_frequent\n",
      "['mirai' 'mirai' 'mirai' ... 'mirai' 'mirai' 'mirai']\n",
      "0.6065904532062935\n",
      "----------------------xxxxxxx----------------------\n",
      "stratified\n",
      "['mirai' 'gafgyt' 'mirai' ... 'gafgyt' 'mirai' 'gafgyt']\n",
      "0.5036631041183477\n",
      "----------------------xxxxxxx----------------------\n",
      "constant\n",
      "['benign' 'benign' 'benign' ... 'benign' 'benign' 'benign']\n",
      "0.02313828120102802\n",
      "----------------------xxxxxxx----------------------\n",
      "uniform\n",
      "['gafgyt' 'mirai' 'benign' ... 'benign' 'gafgyt' 'gafgyt']\n",
      "0.33350310286466495\n",
      "----------------------xxxxxxx----------------------\n"
     ]
    }
   ],
   "source": [
    "# with scaling\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dummy_classifier(dataframe[k], True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "| Device | Most Frequent | Stratified | Constant | Uniform |\n",
    "|---|---|---|---|---|\n",
    "| Danmin Door Bell  | 0.64 | 0.51 | 0.049 | 0.33 |\n",
    "| Ecobee_Thermostat | 0.61  | 0.51 | 0.016 | 0.33 |\n",
    "| Ennio_Doorbell | 0.89 | 0.80 | 0.11 | 0.50 |\n",
    "| Philips_B120N10_Baby_Monitor | 0.55 | 0.41 | 0.15 | 0.33 |\n",
    "| Provision_PT_737E_Security_Camera | 0.52 | 0.44 | 0.075  | 0.33 |\n",
    "| Provision_PT_838_Security_Camera | 0.51  | 0.41 | 0.11 | 0.33  |\n",
    "| Samsung_SNH_1011_N_Webcam | 0.86 | 0.76 | 0.13 | 0.49  |\n",
    "| SimpleHome_XCS7_1002_WHT_Security_Camera | 0.59 | 0.48 | 0.054 | 0.33 |\n",
    "| SimpleHome_XCS7_1003_WHT_Security_Camera | 0.60  | 0.50 | 0.023 | 0.33 |"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Random Forest Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def rf_classifier(data, device_name, scaling=False,):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        model_name = f'{device_name}_without_scaling_unbalanced_model.pkl'\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        model_name = f'{device_name}_with_scaling_unbalanced_model.pkl'\n",
    "    \n",
    "    clf = RandomForestClassifier()\n",
    "    model_res = clf.fit(X_train, y_train)\n",
    "    y_pred = model_res.predict(X_test)\n",
    "    y_pred_prob = model_res.predict_proba(X_test)\n",
    "    lr_probs = y_pred_prob[:,1]\n",
    "    ac = accuracy_score(y_test, y_pred)\n",
    "    importances = pd.DataFrame({'feature':X.columns,'importance':np.round(clf.feature_importances_,3)})\n",
    "    importances = importances.sort_values('importance',ascending=False).set_index('feature')\n",
    "\n",
    "    f1 = f1_score(y_test, y_pred, average='weighted')\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "    print(importances.head(20))\n",
    "    \n",
    "    print('Random Forest: Accuracy=%.3f' % (ac))\n",
    "\n",
    "    print('Random Forest: f1-score=%.3f' % (f1))\n",
    "    print('Random Forest: Confusion Matrix', cm)\n",
    "    print('Random Forest : Classification Report',classification_report(y_test, y_pred))\n",
    "    \n",
    "    best_model = clf\n",
    "    best_model.version = 1.0\n",
    "    best_model.pandas_version = pd.__version__\n",
    "    best_model.numpy_version = np.__version__\n",
    "    best_model.sklearn_version = sklearn_version\n",
    "    #best_model.X_columns = [col for col in X_train.columns]\n",
    "    best_model.build_datetime = datetime.now()\n",
    "    \n",
    "    modelpath = 'models'\n",
    "    if not os.path.exists(modelpath):\n",
    "        os.mkdir(modelpath)\n",
    "    iotmodel_path = os.path.join(modelpath, model_name)\n",
    "    if not os.path.exists(iotmodel_path):\n",
    "        with open(iotmodel_path, 'wb') as f:\n",
    "            pickle.dump(best_model, f)\n",
    "    \n",
    "    print('---------------------- Done ---------------------')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                     importance\n",
      "feature                        \n",
      "MI_dir_L0.01_weight       0.067\n",
      "H_L0.01_weight            0.054\n",
      "MI_dir_L0.01_mean         0.052\n",
      "MI_dir_L0.1_weight        0.050\n",
      "H_L0.1_weight             0.048\n",
      "MI_dir_L3_weight          0.046\n",
      "MI_dir_L1_weight          0.039\n",
      "H_L3_weight               0.035\n",
      "H_L1_weight               0.028\n",
      "MI_dir_L5_weight          0.027\n",
      "HH_L1_weight              0.025\n",
      "HH_jit_L1_weight          0.024\n",
      "HH_jit_L3_mean            0.023\n",
      "HH_jit_L3_weight          0.023\n",
      "MI_dir_L5_variance        0.023\n",
      "H_L5_weight               0.021\n",
      "H_L0.01_mean              0.021\n",
      "HH_L3_weight              0.018\n",
      "HH_L0.1_weight            0.017\n",
      "H_L1_mean                 0.017\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 15000      3      0]\n",
      " [     6  94853      0]\n",
      " [     0      0 195628]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15003\n",
      "      gafgyt       1.00      1.00      1.00     94859\n",
      "       mirai       1.00      1.00      1.00    195628\n",
      "\n",
      "    accuracy                           1.00    305490\n",
      "   macro avg       1.00      1.00      1.00    305490\n",
      "weighted avg       1.00      1.00      1.00    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                     importance\n",
      "feature                        \n",
      "MI_dir_L3_variance        0.074\n",
      "MI_dir_L0.1_weight        0.065\n",
      "MI_dir_L5_variance        0.056\n",
      "H_L3_variance             0.054\n",
      "H_L1_weight               0.045\n",
      "H_L5_variance             0.044\n",
      "H_L0.1_weight             0.042\n",
      "H_L1_variance             0.042\n",
      "MI_dir_L0.01_weight       0.031\n",
      "MI_dir_L1_weight          0.029\n",
      "MI_dir_L0.01_mean         0.026\n",
      "H_L0.01_mean              0.025\n",
      "H_L3_mean                 0.025\n",
      "MI_dir_L1_mean            0.022\n",
      "HH_jit_L5_mean            0.018\n",
      "MI_dir_L1_variance        0.018\n",
      "H_L0.1_mean               0.017\n",
      "MI_dir_L0.1_mean          0.017\n",
      "HH_jit_L0.1_mean          0.016\n",
      "HH_jit_L1_weight          0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[  4016      0      0]\n",
      " [     0  92806      0]\n",
      " [     2      0 153939]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      4016\n",
      "      gafgyt       1.00      1.00      1.00     92806\n",
      "       mirai       1.00      1.00      1.00    153941\n",
      "\n",
      "    accuracy                           1.00    250763\n",
      "   macro avg       1.00      1.00      1.00    250763\n",
      "weighted avg       1.00      1.00      1.00    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                    importance\n",
      "feature                       \n",
      "HH_jit_L0.1_mean         0.107\n",
      "HH_L0.01_radius          0.102\n",
      "MI_dir_L0.01_mean        0.097\n",
      "HH_jit_L0.01_mean        0.093\n",
      "H_L0.01_mean             0.065\n",
      "HH_L0.01_mean            0.056\n",
      "HH_L0.1_mean             0.054\n",
      "HH_L0.01_std             0.037\n",
      "HpHp_L3_mean             0.033\n",
      "H_L0.1_mean              0.030\n",
      "HpHp_L1_mean             0.021\n",
      "HpHp_L0.01_weight        0.020\n",
      "HpHp_L0.1_mean           0.019\n",
      "HH_L0.01_magnitude       0.017\n",
      "HpHp_L5_mean             0.014\n",
      "HH_L1_mean               0.014\n",
      "HH_L5_mean               0.014\n",
      "HpHp_L0.01_mean          0.014\n",
      "HH_L3_mean               0.013\n",
      "HpHp_L0.01_std           0.012\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[11903     0]\n",
      " [    0 94747]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "      gafgyt       1.00      1.00      1.00     94747\n",
      "\n",
      "    accuracy                           1.00    106650\n",
      "   macro avg       1.00      1.00      1.00    106650\n",
      "weighted avg       1.00      1.00      1.00    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                     importance\n",
      "feature                        \n",
      "H_L0.1_weight             0.072\n",
      "MI_dir_L0.01_weight       0.058\n",
      "MI_dir_L3_weight          0.050\n",
      "MI_dir_L1_weight          0.048\n",
      "H_L0.01_weight            0.043\n",
      "H_L1_weight               0.040\n",
      "MI_dir_L5_weight          0.035\n",
      "MI_dir_L0.1_weight        0.027\n",
      "H_L5_weight               0.025\n",
      "HH_jit_L0.01_mean         0.022\n",
      "HH_L0.01_magnitude        0.020\n",
      "HH_L1_weight              0.020\n",
      "HH_jit_L3_mean            0.019\n",
      "H_L3_weight               0.019\n",
      "HH_jit_L3_weight          0.019\n",
      "HH_L5_weight              0.019\n",
      "HpHp_L1_magnitude         0.019\n",
      "MI_dir_L0.1_mean          0.019\n",
      "HpHp_L0.01_weight         0.018\n",
      "HH_L3_weight              0.018\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 52496      2      0]\n",
      " [     4  93876      0]\n",
      " [     0      0 183226]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52498\n",
      "      gafgyt       1.00      1.00      1.00     93880\n",
      "       mirai       1.00      1.00      1.00    183226\n",
      "\n",
      "    accuracy                           1.00    329604\n",
      "   macro avg       1.00      1.00      1.00    329604\n",
      "weighted avg       1.00      1.00      1.00    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.1_weight          0.112\n",
      "H_L0.1_weight               0.077\n",
      "H_L0.01_weight              0.051\n",
      "MI_dir_L1_weight            0.049\n",
      "MI_dir_L0.01_weight         0.045\n",
      "HH_L0.1_weight              0.037\n",
      "H_L1_weight                 0.031\n",
      "H_L0.01_variance            0.030\n",
      "H_L0.1_variance             0.030\n",
      "MI_dir_L0.1_mean            0.027\n",
      "HH_jit_L0.1_weight          0.026\n",
      "HH_jit_L3_mean              0.025\n",
      "HH_jit_L1_weight            0.021\n",
      "HH_jit_L3_weight            0.021\n",
      "MI_dir_L0.1_variance        0.021\n",
      "HH_L3_weight                0.019\n",
      "MI_dir_L3_weight            0.018\n",
      "MI_dir_L0.01_variance       0.017\n",
      "H_L0.1_mean                 0.016\n",
      "HH_jit_L0.01_mean           0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 18758      0      0]\n",
      " [     2  98604      0]\n",
      " [     1      0 131113]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18758\n",
      "      gafgyt       1.00      1.00      1.00     98606\n",
      "       mirai       1.00      1.00      1.00    131114\n",
      "\n",
      "    accuracy                           1.00    248478\n",
      "   macro avg       1.00      1.00      1.00    248478\n",
      "weighted avg       1.00      1.00      1.00    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.065\n",
      "H_L1_weight                 0.062\n",
      "H_L0.01_weight              0.059\n",
      "MI_dir_L0.1_weight          0.054\n",
      "MI_dir_L0.01_weight         0.054\n",
      "HH_L0.1_weight              0.034\n",
      "MI_dir_L1_weight            0.034\n",
      "HH_L1_weight                0.030\n",
      "MI_dir_L3_weight            0.029\n",
      "H_L0.01_mean                0.029\n",
      "H_L0.1_mean                 0.026\n",
      "HH_jit_L5_mean              0.026\n",
      "H_L0.01_variance            0.025\n",
      "H_L3_weight                 0.025\n",
      "HH_L3_weight                0.024\n",
      "HH_jit_L0.1_mean            0.024\n",
      "HH_jit_L0.1_weight          0.021\n",
      "MI_dir_L0.01_variance       0.020\n",
      "HH_L5_weight                0.018\n",
      "MI_dir_L5_weight            0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 29698      0      0]\n",
      " [     1  92246      0]\n",
      " [     0      0 129123]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29698\n",
      "      gafgyt       1.00      1.00      1.00     92247\n",
      "       mirai       1.00      1.00      1.00    129123\n",
      "\n",
      "    accuracy                           1.00    251068\n",
      "   macro avg       1.00      1.00      1.00    251068\n",
      "weighted avg       1.00      1.00      1.00    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.1_mean            0.178\n",
      "MI_dir_L0.01_variance       0.091\n",
      "MI_dir_L0.01_mean           0.087\n",
      "H_L0.01_variance            0.075\n",
      "HH_jit_L0.1_mean            0.072\n",
      "MI_dir_L0.1_variance        0.069\n",
      "HH_L0.01_radius             0.061\n",
      "H_L0.1_mean                 0.048\n",
      "H_L0.01_mean                0.043\n",
      "HH_jit_L0.01_mean           0.040\n",
      "H_L0.1_variance             0.030\n",
      "MI_dir_L1_mean              0.027\n",
      "HpHp_L0.01_magnitude        0.025\n",
      "HH_L0.01_std                0.017\n",
      "HpHp_L3_mean                0.010\n",
      "HpHp_L0.1_mean              0.010\n",
      "HpHp_L1_magnitude           0.009\n",
      "HpHp_L0.01_weight           0.009\n",
      "HH_L5_mean                  0.009\n",
      "HpHp_L1_mean                0.009\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15682     2]\n",
      " [    5 96878]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "      gafgyt       1.00      1.00      1.00     96883\n",
      "\n",
      "    accuracy                           1.00    112567\n",
      "   macro avg       1.00      1.00      1.00    112567\n",
      "weighted avg       1.00      1.00      1.00    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L5_variance          0.071\n",
      "H_L0.1_weight               0.062\n",
      "MI_dir_L0.1_weight          0.057\n",
      "MI_dir_L3_variance          0.047\n",
      "MI_dir_L1_weight            0.042\n",
      "H_L1_weight                 0.034\n",
      "MI_dir_L0.01_mean           0.034\n",
      "H_L0.01_mean                0.033\n",
      "MI_dir_L0.01_variance       0.030\n",
      "MI_dir_L0.01_weight         0.028\n",
      "H_L3_variance               0.027\n",
      "H_L3_weight                 0.026\n",
      "MI_dir_L0.1_mean            0.023\n",
      "HH_jit_L1_weight            0.022\n",
      "H_L5_weight                 0.021\n",
      "H_L5_variance               0.020\n",
      "H_L0.1_mean                 0.019\n",
      "MI_dir_L0.1_variance        0.018\n",
      "HH_jit_L1_mean              0.018\n",
      "HH_jit_L3_mean              0.017\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 14055      0      0]\n",
      " [     0  90533      0]\n",
      " [     1      0 154328]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14055\n",
      "      gafgyt       1.00      1.00      1.00     90533\n",
      "       mirai       1.00      1.00      1.00    154329\n",
      "\n",
      "    accuracy                           1.00    258917\n",
      "   macro avg       1.00      1.00      1.00    258917\n",
      "weighted avg       1.00      1.00      1.00    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                     importance\n",
      "feature                        \n",
      "H_L3_variance             0.069\n",
      "MI_dir_L5_variance        0.057\n",
      "H_L5_variance             0.043\n",
      "H_L0.1_weight             0.041\n",
      "MI_dir_L3_variance        0.040\n",
      "H_L1_weight               0.039\n",
      "H_L1_variance             0.037\n",
      "MI_dir_L0.1_weight        0.036\n",
      "H_L0.1_mean               0.036\n",
      "MI_dir_L0.1_mean          0.034\n",
      "MI_dir_L0.01_mean         0.034\n",
      "H_L3_weight               0.033\n",
      "MI_dir_L1_weight          0.029\n",
      "MI_dir_L0.01_weight       0.024\n",
      "H_L0.01_mean              0.024\n",
      "H_L0.01_variance          0.024\n",
      "H_L1_mean                 0.022\n",
      "H_L0.01_weight            0.018\n",
      "HH_jit_L0.01_mean         0.017\n",
      "HH_jit_L3_mean            0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[  5905      1      0]\n",
      " [     0  94511      0]\n",
      " [     0      1 154830]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5906\n",
      "      gafgyt       1.00      1.00      1.00     94511\n",
      "       mirai       1.00      1.00      1.00    154831\n",
      "\n",
      "    accuracy                           1.00    255248\n",
      "   macro avg       1.00      1.00      1.00    255248\n",
      "weighted avg       1.00      1.00      1.00    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with out scaling\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier(dataframe[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                     importance\n",
      "feature                        \n",
      "H_L0.1_weight             0.076\n",
      "H_L0.01_weight            0.069\n",
      "H_L1_weight               0.062\n",
      "MI_dir_L0.01_weight       0.058\n",
      "MI_dir_L0.1_weight        0.050\n",
      "MI_dir_L1_weight          0.046\n",
      "MI_dir_L0.01_mean         0.038\n",
      "MI_dir_L3_weight          0.036\n",
      "H_L0.1_mean               0.030\n",
      "HH_L3_weight              0.029\n",
      "HH_jit_L3_weight          0.025\n",
      "HH_L0.01_magnitude        0.024\n",
      "H_L0.01_mean              0.021\n",
      "H_L3_weight               0.020\n",
      "MI_dir_L5_weight          0.019\n",
      "MI_dir_L0.1_mean          0.019\n",
      "HH_L0.1_weight            0.018\n",
      "H_L5_weight               0.017\n",
      "HH_L1_weight              0.016\n",
      "HH_jit_L0.1_weight        0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 15002      1      0]\n",
      " [     9  94850      0]\n",
      " [     0      0 195628]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15003\n",
      "      gafgyt       1.00      1.00      1.00     94859\n",
      "       mirai       1.00      1.00      1.00    195628\n",
      "\n",
      "    accuracy                           1.00    305490\n",
      "   macro avg       1.00      1.00      1.00    305490\n",
      "weighted avg       1.00      1.00      1.00    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                     importance\n",
      "feature                        \n",
      "MI_dir_L5_variance        0.062\n",
      "H_L3_variance             0.061\n",
      "H_L0.1_weight             0.059\n",
      "H_L1_weight               0.048\n",
      "H_L5_variance             0.047\n",
      "MI_dir_L3_variance        0.047\n",
      "H_L0.01_mean              0.038\n",
      "H_L0.01_weight            0.037\n",
      "H_L1_variance             0.035\n",
      "MI_dir_L1_weight          0.035\n",
      "MI_dir_L0.1_weight        0.034\n",
      "MI_dir_L0.01_weight       0.030\n",
      "MI_dir_L0.01_mean         0.029\n",
      "H_L0.1_mean               0.025\n",
      "HH_jit_L1_weight          0.025\n",
      "HH_L0.1_weight            0.022\n",
      "MI_dir_L0.1_mean          0.020\n",
      "H_L3_weight               0.018\n",
      "HH_L0.01_magnitude        0.018\n",
      "MI_dir_L3_weight          0.017\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[  4016      0      0]\n",
      " [     0  92806      0]\n",
      " [     2      0 153939]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      4016\n",
      "      gafgyt       1.00      1.00      1.00     92806\n",
      "       mirai       1.00      1.00      1.00    153941\n",
      "\n",
      "    accuracy                           1.00    250763\n",
      "   macro avg       1.00      1.00      1.00    250763\n",
      "weighted avg       1.00      1.00      1.00    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                      importance\n",
      "feature                         \n",
      "HH_jit_L0.01_mean          0.186\n",
      "MI_dir_L0.01_mean          0.092\n",
      "H_L0.01_mean               0.083\n",
      "HpHp_L0.01_weight          0.067\n",
      "HH_L0.01_mean              0.066\n",
      "HH_L0.01_radius            0.047\n",
      "MI_dir_L0.1_mean           0.038\n",
      "HH_L0.01_magnitude         0.033\n",
      "HpHp_L0.1_mean             0.032\n",
      "HpHp_L0.01_mean            0.028\n",
      "HpHp_L1_mean               0.026\n",
      "HpHp_L3_mean               0.024\n",
      "HH_L0.1_mean               0.021\n",
      "HH_L0.1_std                0.017\n",
      "HH_L5_mean                 0.014\n",
      "HH_L5_magnitude            0.013\n",
      "HH_L0.01_std               0.013\n",
      "HpHp_L0.01_magnitude       0.012\n",
      "HH_L1_magnitude            0.012\n",
      "HH_L0.1_radius             0.011\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[11903     0]\n",
      " [    0 94747]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "      gafgyt       1.00      1.00      1.00     94747\n",
      "\n",
      "    accuracy                           1.00    106650\n",
      "   macro avg       1.00      1.00      1.00    106650\n",
      "weighted avg       1.00      1.00      1.00    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.1_weight          0.070\n",
      "MI_dir_L0.01_weight         0.060\n",
      "H_L0.1_weight               0.059\n",
      "H_L0.01_weight              0.045\n",
      "MI_dir_L3_weight            0.039\n",
      "H_L5_weight                 0.032\n",
      "MI_dir_L1_weight            0.030\n",
      "H_L3_weight                 0.028\n",
      "HpHp_L0.01_std              0.026\n",
      "HpHp_L0.01_radius           0.024\n",
      "H_L1_weight                 0.023\n",
      "HH_L0.01_magnitude          0.022\n",
      "HH_jit_L1_weight            0.021\n",
      "H_L0.01_mean                0.021\n",
      "HH_L0.01_radius             0.020\n",
      "HH_L0.1_magnitude           0.019\n",
      "HH_L3_weight                0.019\n",
      "HH_L0.1_weight              0.018\n",
      "HH_jit_L0.1_weight          0.018\n",
      "MI_dir_L0.01_variance       0.017\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 52496      2      0]\n",
      " [     2  93878      0]\n",
      " [     0      0 183226]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52498\n",
      "      gafgyt       1.00      1.00      1.00     93880\n",
      "       mirai       1.00      1.00      1.00    183226\n",
      "\n",
      "    accuracy                           1.00    329604\n",
      "   macro avg       1.00      1.00      1.00    329604\n",
      "weighted avg       1.00      1.00      1.00    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                     importance\n",
      "feature                        \n",
      "H_L0.01_weight            0.073\n",
      "H_L0.1_weight             0.071\n",
      "MI_dir_L0.1_weight        0.070\n",
      "H_L1_weight               0.047\n",
      "H_L0.1_mean               0.037\n",
      "MI_dir_L1_weight          0.034\n",
      "H_L0.01_variance          0.034\n",
      "MI_dir_L0.01_weight       0.030\n",
      "H_L0.01_mean              0.029\n",
      "HH_L1_weight              0.029\n",
      "H_L5_weight               0.028\n",
      "HH_jit_L0.1_weight        0.025\n",
      "MI_dir_L0.1_mean          0.025\n",
      "HH_jit_L3_weight          0.023\n",
      "MI_dir_L1_mean            0.023\n",
      "MI_dir_L0.01_mean         0.022\n",
      "HH_L0.1_weight            0.021\n",
      "H_L3_weight               0.021\n",
      "H_L1_mean                 0.019\n",
      "H_L0.1_variance           0.019\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 18758      0      0]\n",
      " [     2  98604      0]\n",
      " [     1      0 131113]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18758\n",
      "      gafgyt       1.00      1.00      1.00     98606\n",
      "       mirai       1.00      1.00      1.00    131114\n",
      "\n",
      "    accuracy                           1.00    248478\n",
      "   macro avg       1.00      1.00      1.00    248478\n",
      "weighted avg       1.00      1.00      1.00    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.1_weight          0.084\n",
      "MI_dir_L0.01_weight         0.071\n",
      "H_L0.1_weight               0.070\n",
      "H_L1_weight                 0.056\n",
      "H_L0.01_mean                0.035\n",
      "H_L3_weight                 0.035\n",
      "MI_dir_L1_weight            0.032\n",
      "H_L0.1_mean                 0.027\n",
      "MI_dir_L0.01_mean           0.025\n",
      "MI_dir_L3_weight            0.024\n",
      "HH_jit_L1_weight            0.024\n",
      "MI_dir_L0.1_mean            0.023\n",
      "HH_jit_L0.1_weight          0.023\n",
      "HH_L0.1_weight              0.022\n",
      "H_L0.01_weight              0.022\n",
      "HH_L3_weight                0.021\n",
      "HH_L5_weight                0.020\n",
      "HH_jit_L0.01_mean           0.018\n",
      "MI_dir_L0.01_variance       0.018\n",
      "HH_L0.01_weight             0.017\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 29698      0      0]\n",
      " [     2  92245      0]\n",
      " [     0      0 129123]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29698\n",
      "      gafgyt       1.00      1.00      1.00     92247\n",
      "       mirai       1.00      1.00      1.00    129123\n",
      "\n",
      "    accuracy                           1.00    251068\n",
      "   macro avg       1.00      1.00      1.00    251068\n",
      "weighted avg       1.00      1.00      1.00    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_mean                 0.139\n",
      "H_L0.1_variance             0.091\n",
      "H_L0.01_mean                0.082\n",
      "MI_dir_L0.1_mean            0.082\n",
      "H_L0.01_variance            0.081\n",
      "MI_dir_L0.01_mean           0.062\n",
      "MI_dir_L0.01_variance       0.051\n",
      "HH_jit_L0.01_mean           0.048\n",
      "HH_L0.01_radius             0.043\n",
      "HpHp_L0.01_weight           0.037\n",
      "H_L1_mean                   0.035\n",
      "HpHp_L0.1_mean              0.028\n",
      "HH_L0.1_magnitude           0.025\n",
      "HH_L0.01_std                0.025\n",
      "MI_dir_L0.1_variance        0.020\n",
      "HH_L0.01_magnitude          0.017\n",
      "HH_L0.01_mean               0.017\n",
      "HpHp_L3_magnitude           0.017\n",
      "HH_L5_magnitude             0.017\n",
      "HpHp_L0.01_mean             0.009\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15682     2]\n",
      " [    7 96876]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "      gafgyt       1.00      1.00      1.00     96883\n",
      "\n",
      "    accuracy                           1.00    112567\n",
      "   macro avg       1.00      1.00      1.00    112567\n",
      "weighted avg       1.00      1.00      1.00    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.068\n",
      "H_L3_variance               0.060\n",
      "MI_dir_L0.1_weight          0.052\n",
      "MI_dir_L3_variance          0.051\n",
      "H_L5_variance               0.041\n",
      "MI_dir_L1_weight            0.038\n",
      "H_L0.01_mean                0.037\n",
      "MI_dir_L5_variance          0.032\n",
      "MI_dir_L0.01_mean           0.030\n",
      "MI_dir_L0.1_mean            0.027\n",
      "H_L0.1_mean                 0.027\n",
      "MI_dir_L0.01_weight         0.026\n",
      "H_L1_weight                 0.026\n",
      "H_L1_variance               0.024\n",
      "H_L0.01_weight              0.023\n",
      "MI_dir_L3_weight            0.021\n",
      "MI_dir_L0.1_variance        0.019\n",
      "MI_dir_L0.01_variance       0.019\n",
      "H_L0.01_variance            0.018\n",
      "HH_L0.1_weight              0.017\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 14055      0      0]\n",
      " [     1  90532      0]\n",
      " [     1      0 154328]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14055\n",
      "      gafgyt       1.00      1.00      1.00     90533\n",
      "       mirai       1.00      1.00      1.00    154329\n",
      "\n",
      "    accuracy                           1.00    258917\n",
      "   macro avg       1.00      1.00      1.00    258917\n",
      "weighted avg       1.00      1.00      1.00    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L5_variance          0.069\n",
      "MI_dir_L3_variance          0.065\n",
      "H_L3_variance               0.059\n",
      "H_L0.1_weight               0.057\n",
      "H_L5_variance               0.046\n",
      "H_L1_weight                 0.043\n",
      "MI_dir_L0.01_weight         0.036\n",
      "H_L0.01_mean                0.031\n",
      "MI_dir_L3_weight            0.030\n",
      "MI_dir_L0.01_mean           0.029\n",
      "H_L0.01_weight              0.027\n",
      "H_L0.1_mean                 0.026\n",
      "MI_dir_L0.1_weight          0.025\n",
      "MI_dir_L0.01_variance       0.022\n",
      "MI_dir_L1_variance          0.021\n",
      "MI_dir_L1_weight            0.021\n",
      "MI_dir_L1_mean              0.019\n",
      "MI_dir_L0.1_mean            0.019\n",
      "H_L1_variance               0.015\n",
      "MI_dir_L5_weight            0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[  5905      1      0]\n",
      " [     0  94511      0]\n",
      " [     2      0 154829]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5906\n",
      "      gafgyt       1.00      1.00      1.00     94511\n",
      "       mirai       1.00      1.00      1.00    154831\n",
      "\n",
      "    accuracy                           1.00    255248\n",
      "   macro avg       1.00      1.00      1.00    255248\n",
      "weighted avg       1.00      1.00      1.00    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with scaling\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier(dataframe[k], k, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def rf_classifier_balanced(data, device_name, scaling=False):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "    nm = NearMiss()\n",
    "    \n",
    "    X_res, y_res = nm.fit_resample(X, y)\n",
    "    print(f'Imblearn Shape:{X_res.shape}, {y_res.shape}')\n",
    "    \n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.30, random_state=47)\n",
    "        model_name = f'{device_name}_without_scaling_balanced_model.pkl'\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        model_name = f'{device_name}_with_scaling_balanced_model.pkl'\n",
    "    \n",
    "    clf = RandomForestClassifier()\n",
    "    model_res = clf.fit(X_train, y_train)\n",
    "    y_pred = model_res.predict(X_test)\n",
    "    y_pred_prob = model_res.predict_proba(X_test)\n",
    "    lr_probs = y_pred_prob[:,1]\n",
    "    ac = accuracy_score(y_test, y_pred)\n",
    "    importances = pd.DataFrame({'feature':X.columns,'importance':np.round(clf.feature_importances_,3)})\n",
    "    importances = importances.sort_values('importance',ascending=False).set_index('feature')\n",
    "\n",
    "    f1 = f1_score(y_test, y_pred, average='weighted')\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "    print(importances.head(20))\n",
    "    \n",
    "    print('Random Forest: Accuracy=%.3f' % (ac))\n",
    "\n",
    "    print('Random Forest: f1-score=%.3f' % (f1))\n",
    "    print('Random Forest: Confusion Matrix', cm)\n",
    "    print('Random Forest : Classification Report',classification_report(y_test, y_pred))\n",
    "    \n",
    "    best_model = clf\n",
    "    best_model.version = 1.0\n",
    "    best_model.pandas_version = pd.__version__\n",
    "    best_model.numpy_version = np.__version__\n",
    "    best_model.sklearn_version = sklearn_version\n",
    "    #best_model.X_columns = [col for col in X_train.columns]\n",
    "    best_model.build_datetime = datetime.now()\n",
    "    \n",
    "    modelpath = 'models'\n",
    "    if not os.path.exists(modelpath):\n",
    "        os.mkdir(modelpath)\n",
    "    iotmodel_path = os.path.join(modelpath, model_name)\n",
    "    if not os.path.exists(iotmodel_path):\n",
    "        with open(iotmodel_path, 'wb') as f:\n",
    "            pickle.dump(best_model, f)\n",
    "    \n",
    "    print('---------------------- Done ---------------------')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(148644, 115), (148644,)\n",
      "                      importance\n",
      "feature                         \n",
      "HH_jit_L0.01_mean          0.097\n",
      "HH_jit_L1_mean             0.075\n",
      "HH_jit_L0.1_mean           0.075\n",
      "H_L0.01_mean               0.072\n",
      "HH_jit_L3_mean             0.070\n",
      "HH_jit_L5_mean             0.068\n",
      "MI_dir_L0.1_weight         0.035\n",
      "HH_L0.01_magnitude         0.031\n",
      "H_L0.1_weight              0.030\n",
      "HpHp_L5_magnitude          0.030\n",
      "MI_dir_L0.01_weight        0.029\n",
      "H_L0.01_weight             0.029\n",
      "HpHp_L0.01_magnitude       0.025\n",
      "HH_L0.1_magnitude          0.025\n",
      "HH_jit_L0.01_weight        0.025\n",
      "MI_dir_L0.01_mean          0.024\n",
      "HpHp_L1_magnitude          0.022\n",
      "HH_L1_magnitude            0.017\n",
      "H_L1_weight                0.017\n",
      "HpHp_L3_magnitude          0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[14849     0     0]\n",
      " [    0 14860     0]\n",
      " [    1     0 14884]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14849\n",
      "      gafgyt       1.00      1.00      1.00     14860\n",
      "       mirai       1.00      1.00      1.00     14885\n",
      "\n",
      "    accuracy                           1.00     44594\n",
      "   macro avg       1.00      1.00      1.00     44594\n",
      "weighted avg       1.00      1.00      1.00     44594\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(39339, 115), (39339,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.098\n",
      "HH_jit_L5_mean              0.094\n",
      "HH_jit_L1_mean              0.086\n",
      "HH_jit_L3_mean              0.083\n",
      "HH_jit_L0.01_mean           0.082\n",
      "H_L0.01_mean                0.051\n",
      "MI_dir_L0.1_mean            0.040\n",
      "H_L0.1_weight               0.034\n",
      "HH_L0.1_weight              0.033\n",
      "HH_jit_L0.01_variance       0.030\n",
      "HpHp_L0.01_weight           0.030\n",
      "MI_dir_L0.1_weight          0.027\n",
      "MI_dir_L0.01_mean           0.027\n",
      "HH_L3_magnitude             0.023\n",
      "HH_jit_L0.01_weight         0.020\n",
      "HpHp_L0.1_weight            0.019\n",
      "H_L0.1_mean                 0.019\n",
      "HH_L0.01_weight             0.015\n",
      "MI_dir_L5_mean              0.013\n",
      "H_L1_mean                   0.012\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[3913    0    0]\n",
      " [   0 3891    0]\n",
      " [   0    0 3998]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      3913\n",
      "      gafgyt       1.00      1.00      1.00      3891\n",
      "       mirai       1.00      1.00      1.00      3998\n",
      "\n",
      "    accuracy                           1.00     11802\n",
      "   macro avg       1.00      1.00      1.00     11802\n",
      "weighted avg       1.00      1.00      1.00     11802\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(78200, 115), (78200,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L3_mean              0.104\n",
      "HH_jit_L0.01_mean           0.103\n",
      "HH_jit_L1_mean              0.063\n",
      "HH_L0.01_magnitude          0.060\n",
      "HH_jit_L0.01_variance       0.060\n",
      "HH_L0.01_weight             0.060\n",
      "HH_jit_L0.1_mean            0.053\n",
      "HH_jit_L0.01_weight         0.050\n",
      "HH_L5_magnitude             0.049\n",
      "HH_L0.1_magnitude           0.049\n",
      "HH_jit_L5_mean              0.041\n",
      "HH_L3_magnitude             0.039\n",
      "HH_jit_L0.1_weight          0.030\n",
      "HH_L1_magnitude             0.029\n",
      "MI_dir_L0.1_mean            0.028\n",
      "HH_jit_L0.1_variance        0.025\n",
      "HH_L0.01_std                0.019\n",
      "H_L0.1_mean                 0.018\n",
      "MI_dir_L1_mean              0.017\n",
      "HH_L0.1_weight              0.012\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[11788     0]\n",
      " [    0 11672]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11788\n",
      "      gafgyt       1.00      1.00      1.00     11672\n",
      "\n",
      "    accuracy                           1.00     23460\n",
      "   macro avg       1.00      1.00      1.00     23460\n",
      "weighted avg       1.00      1.00      1.00     23460\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(525720, 115), (525720,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean              0.086\n",
      "HH_jit_L5_mean              0.075\n",
      "HH_jit_L0.1_mean            0.073\n",
      "HH_jit_L0.01_mean           0.067\n",
      "MI_dir_L0.1_weight          0.067\n",
      "HH_jit_L3_mean              0.060\n",
      "H_L0.1_weight               0.045\n",
      "HH_L0.01_magnitude          0.040\n",
      "HH_jit_L0.01_weight         0.035\n",
      "HH_L0.1_weight              0.035\n",
      "HpHp_L5_magnitude           0.026\n",
      "HH_L1_magnitude             0.026\n",
      "HH_L0.1_magnitude           0.026\n",
      "HH_jit_L0.01_variance       0.025\n",
      "MI_dir_L5_weight            0.024\n",
      "H_L5_weight                 0.024\n",
      "HpHp_L0.01_magnitude        0.021\n",
      "HpHp_L3_magnitude           0.020\n",
      "HH_L5_magnitude             0.015\n",
      "HpHp_L0.1_magnitude         0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[52311     0     0]\n",
      " [    0 52769     0]\n",
      " [    2     0 52634]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52311\n",
      "      gafgyt       1.00      1.00      1.00     52769\n",
      "       mirai       1.00      1.00      1.00     52636\n",
      "\n",
      "    accuracy                           1.00    157716\n",
      "   macro avg       1.00      1.00      1.00    157716\n",
      "weighted avg       1.00      1.00      1.00    157716\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(186462, 115), (186462,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L3_mean              0.107\n",
      "HH_jit_L5_mean              0.087\n",
      "HH_jit_L0.1_mean            0.077\n",
      "HH_jit_L1_mean              0.071\n",
      "MI_dir_L0.1_weight          0.068\n",
      "H_L0.1_weight               0.058\n",
      "HH_jit_L0.01_mean           0.057\n",
      "HH_jit_L0.1_variance        0.044\n",
      "H_L0.01_weight              0.038\n",
      "HH_jit_L0.01_variance       0.035\n",
      "MI_dir_L0.01_weight         0.032\n",
      "HH_jit_L0.01_weight         0.026\n",
      "MI_dir_L0.01_mean           0.024\n",
      "H_L1_weight                 0.022\n",
      "H_L0.1_mean                 0.021\n",
      "H_L1_mean                   0.021\n",
      "HH_L0.01_std                0.020\n",
      "MI_dir_L1_weight            0.017\n",
      "H_L0.1_variance             0.016\n",
      "HH_L0.01_weight             0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[18587     0     0]\n",
      " [    0 18646     0]\n",
      " [    0     0 18706]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18587\n",
      "      gafgyt       1.00      1.00      1.00     18646\n",
      "       mirai       1.00      1.00      1.00     18706\n",
      "\n",
      "    accuracy                           1.00     55939\n",
      "   macro avg       1.00      1.00      1.00     55939\n",
      "weighted avg       1.00      1.00      1.00     55939\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(295542, 115), (295542,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.077\n",
      "HH_jit_L0.1_mean            0.074\n",
      "HH_jit_L3_mean              0.074\n",
      "HH_jit_L1_mean              0.073\n",
      "HH_jit_L5_mean              0.072\n",
      "MI_dir_L0.01_weight         0.050\n",
      "HH_jit_L0.01_weight         0.042\n",
      "H_L0.1_weight               0.040\n",
      "MI_dir_L0.1_weight          0.036\n",
      "HH_jit_L0.01_variance       0.036\n",
      "HH_L0.01_weight             0.036\n",
      "HH_jit_L0.1_weight          0.035\n",
      "MI_dir_L1_weight            0.027\n",
      "HH_L0.1_std                 0.027\n",
      "HH_jit_L0.1_variance        0.026\n",
      "HH_L0.01_std                0.024\n",
      "H_L0.01_mean                0.023\n",
      "H_L0.01_weight              0.022\n",
      "H_L0.1_mean                 0.022\n",
      "MI_dir_L0.01_mean           0.020\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[29478     0     0]\n",
      " [    0 29701     0]\n",
      " [    1     0 29483]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29478\n",
      "      gafgyt       1.00      1.00      1.00     29701\n",
      "       mirai       1.00      1.00      1.00     29484\n",
      "\n",
      "    accuracy                           1.00     88663\n",
      "   macro avg       1.00      1.00      1.00     88663\n",
      "weighted avg       1.00      1.00      1.00     88663\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(104300, 115), (104300,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.101\n",
      "MI_dir_L0.1_mean            0.100\n",
      "HH_jit_L5_mean              0.091\n",
      "H_L0.01_mean                0.070\n",
      "H_L0.1_mean                 0.061\n",
      "HH_jit_L0.01_variance       0.061\n",
      "HH_jit_L0.01_mean           0.061\n",
      "HH_jit_L3_mean              0.051\n",
      "HH_jit_L0.1_mean            0.041\n",
      "HH_jit_L1_mean              0.040\n",
      "HH_jit_L0.01_weight         0.040\n",
      "HpHp_L0.01_weight           0.039\n",
      "HpHp_L0.1_weight            0.029\n",
      "HH_jit_L0.1_variance        0.020\n",
      "HH_L0.01_weight             0.020\n",
      "HH_jit_L0.1_weight          0.020\n",
      "HH_L0.1_weight              0.020\n",
      "H_L0.01_variance            0.019\n",
      "HH_L0.01_std                0.019\n",
      "HH_L0.01_magnitude          0.019\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15642     0]\n",
      " [    0 15648]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15642\n",
      "      gafgyt       1.00      1.00      1.00     15648\n",
      "\n",
      "    accuracy                           1.00     31290\n",
      "   macro avg       1.00      1.00      1.00     31290\n",
      "weighted avg       1.00      1.00      1.00     31290\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(139755, 115), (139755,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.107\n",
      "HH_jit_L5_mean              0.093\n",
      "HH_jit_L0.01_mean           0.088\n",
      "HH_jit_L3_mean              0.081\n",
      "HH_jit_L1_mean              0.069\n",
      "HH_L1_magnitude             0.040\n",
      "HH_L0.1_magnitude           0.037\n",
      "HH_L0.01_magnitude          0.033\n",
      "HH_jit_L0.01_weight         0.030\n",
      "MI_dir_L0.01_mean           0.027\n",
      "MI_dir_L0.1_weight          0.025\n",
      "HH_jit_L0.1_variance        0.025\n",
      "HH_L3_magnitude             0.025\n",
      "H_L0.1_weight               0.023\n",
      "MI_dir_L0.1_mean            0.021\n",
      "MI_dir_L1_weight            0.021\n",
      "HH_jit_L0.01_variance       0.020\n",
      "MI_dir_L3_weight            0.016\n",
      "HpHp_L0.1_magnitude         0.016\n",
      "HpHp_L0.01_weight           0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[13999     0     0]\n",
      " [    0 13971     0]\n",
      " [    0     0 13957]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     13999\n",
      "      gafgyt       1.00      1.00      1.00     13971\n",
      "       mirai       1.00      1.00      1.00     13957\n",
      "\n",
      "    accuracy                           1.00     41927\n",
      "   macro avg       1.00      1.00      1.00     41927\n",
      "weighted avg       1.00      1.00      1.00     41927\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(58584, 115), (58584,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean              0.114\n",
      "HH_jit_L5_mean              0.109\n",
      "HH_jit_L0.01_mean           0.088\n",
      "HH_jit_L3_mean              0.068\n",
      "H_L0.01_mean                0.042\n",
      "H_L0.1_mean                 0.041\n",
      "MI_dir_L0.1_mean            0.038\n",
      "MI_dir_L0.01_mean           0.038\n",
      "HH_jit_L0.1_mean            0.037\n",
      "H_L0.1_weight               0.030\n",
      "MI_dir_L0.1_weight          0.028\n",
      "MI_dir_L0.01_variance       0.023\n",
      "H_L0.01_variance            0.022\n",
      "H_L5_weight                 0.021\n",
      "HH_jit_L0.01_weight         0.019\n",
      "H_L0.01_weight              0.019\n",
      "MI_dir_L1_weight            0.017\n",
      "HH_L0.01_weight             0.016\n",
      "H_L3_weight                 0.015\n",
      "MI_dir_L0.1_variance        0.014\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[5743    0    0]\n",
      " [   0 5859    0]\n",
      " [   0    0 5974]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5743\n",
      "      gafgyt       1.00      1.00      1.00      5859\n",
      "       mirai       1.00      1.00      1.00      5974\n",
      "\n",
      "    accuracy                           1.00     17576\n",
      "   macro avg       1.00      1.00      1.00     17576\n",
      "weighted avg       1.00      1.00      1.00     17576\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with out scaling and balanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier_balanced(dataframe[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(148644, 115), (148644,)\n",
      "                    importance\n",
      "feature                       \n",
      "HH_jit_L5_mean           0.097\n",
      "HH_jit_L0.01_mean        0.082\n",
      "HH_jit_L3_mean           0.075\n",
      "HH_jit_L0.1_mean         0.075\n",
      "HH_jit_L1_mean           0.060\n",
      "MI_dir_L0.01_mean        0.058\n",
      "HH_L5_magnitude          0.055\n",
      "MI_dir_L0.1_weight       0.048\n",
      "H_L0.1_weight            0.047\n",
      "HH_L0.1_magnitude        0.040\n",
      "HpHp_L5_magnitude        0.031\n",
      "HH_L3_magnitude          0.026\n",
      "HH_L1_magnitude          0.023\n",
      "MI_dir_L3_weight         0.022\n",
      "H_L0.01_mean             0.020\n",
      "HpHp_L3_magnitude        0.020\n",
      "HpHp_L1_magnitude        0.015\n",
      "HH_L0.01_magnitude       0.015\n",
      "MI_dir_L5_weight         0.015\n",
      "HH_L0.01_mean            0.014\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[14849     0     0]\n",
      " [    0 14860     0]\n",
      " [    1     0 14884]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14849\n",
      "      gafgyt       1.00      1.00      1.00     14860\n",
      "       mirai       1.00      1.00      1.00     14885\n",
      "\n",
      "    accuracy                           1.00     44594\n",
      "   macro avg       1.00      1.00      1.00     44594\n",
      "weighted avg       1.00      1.00      1.00     44594\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(39339, 115), (39339,)\n",
      "                      importance\n",
      "feature                         \n",
      "HH_jit_L1_mean             0.097\n",
      "HH_jit_L5_mean             0.075\n",
      "HH_jit_L0.01_mean          0.072\n",
      "HH_jit_L0.1_mean           0.068\n",
      "HH_jit_L3_mean             0.067\n",
      "HH_jit_L0.01_weight        0.046\n",
      "MI_dir_L0.1_weight         0.044\n",
      "H_L0.01_mean               0.038\n",
      "HpHp_L0.01_weight          0.036\n",
      "MI_dir_L0.01_mean          0.030\n",
      "HH_L0.01_weight            0.028\n",
      "MI_dir_L0.1_mean           0.026\n",
      "HH_L0.01_magnitude         0.023\n",
      "HpHp_L0.1_weight           0.019\n",
      "H_L0.1_mean                0.019\n",
      "H_L0.1_weight              0.017\n",
      "HH_L0.1_weight             0.016\n",
      "HH_L1_magnitude            0.016\n",
      "MI_dir_L5_weight           0.016\n",
      "MI_dir_L0.1_variance       0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[3913    0    0]\n",
      " [   0 3891    0]\n",
      " [   0    0 3998]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      3913\n",
      "      gafgyt       1.00      1.00      1.00      3891\n",
      "       mirai       1.00      1.00      1.00      3998\n",
      "\n",
      "    accuracy                           1.00     11802\n",
      "   macro avg       1.00      1.00      1.00     11802\n",
      "weighted avg       1.00      1.00      1.00     11802\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(78200, 115), (78200,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L3_mean              0.132\n",
      "HH_jit_L0.1_mean            0.092\n",
      "HH_jit_L1_mean              0.086\n",
      "HH_jit_L0.01_weight         0.074\n",
      "HH_jit_L5_mean              0.062\n",
      "HH_L0.01_weight             0.059\n",
      "HH_jit_L0.01_mean           0.050\n",
      "MI_dir_L0.01_weight         0.048\n",
      "HH_jit_L0.1_weight          0.041\n",
      "HH_L5_magnitude             0.038\n",
      "HH_L0.1_magnitude           0.032\n",
      "HH_L3_magnitude             0.030\n",
      "HH_L0.1_weight              0.030\n",
      "MI_dir_L0.1_weight          0.029\n",
      "MI_dir_L0.1_mean            0.027\n",
      "MI_dir_L0.01_mean           0.019\n",
      "H_L0.01_mean                0.019\n",
      "H_L0.1_mean                 0.018\n",
      "H_L0.01_weight              0.018\n",
      "MI_dir_L0.01_variance       0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[11788     0]\n",
      " [    0 11672]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11788\n",
      "      gafgyt       1.00      1.00      1.00     11672\n",
      "\n",
      "    accuracy                           1.00     23460\n",
      "   macro avg       1.00      1.00      1.00     23460\n",
      "weighted avg       1.00      1.00      1.00     23460\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(525720, 115), (525720,)\n",
      "                     importance\n",
      "feature                        \n",
      "HH_jit_L0.1_mean          0.091\n",
      "HH_jit_L0.01_mean         0.090\n",
      "HH_jit_L3_mean            0.078\n",
      "HH_jit_L1_mean            0.074\n",
      "HH_L0.01_magnitude        0.061\n",
      "HH_jit_L5_mean            0.061\n",
      "MI_dir_L0.1_weight        0.056\n",
      "H_L0.1_weight             0.050\n",
      "HH_L0.01_weight           0.030\n",
      "MI_dir_L1_weight          0.029\n",
      "MI_dir_L3_weight          0.027\n",
      "HH_L0.1_weight            0.025\n",
      "HH_jit_L0.1_weight        0.025\n",
      "MI_dir_L0.01_weight       0.025\n",
      "HH_jit_L0.01_weight       0.025\n",
      "HpHp_L3_magnitude         0.020\n",
      "MI_dir_L0.1_mean          0.019\n",
      "HH_L3_magnitude           0.015\n",
      "HH_L5_magnitude           0.015\n",
      "HpHp_L0.1_magnitude       0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[52311     0     0]\n",
      " [    0 52769     0]\n",
      " [    1     0 52635]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52311\n",
      "      gafgyt       1.00      1.00      1.00     52769\n",
      "       mirai       1.00      1.00      1.00     52636\n",
      "\n",
      "    accuracy                           1.00    157716\n",
      "   macro avg       1.00      1.00      1.00    157716\n",
      "weighted avg       1.00      1.00      1.00    157716\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(186462, 115), (186462,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.105\n",
      "HH_jit_L1_mean              0.092\n",
      "HH_jit_L0.01_mean           0.075\n",
      "HH_jit_L3_mean              0.071\n",
      "HH_jit_L0.1_mean            0.059\n",
      "H_L0.1_weight               0.055\n",
      "HH_jit_L0.01_weight         0.051\n",
      "MI_dir_L0.1_weight          0.050\n",
      "H_L0.1_mean                 0.043\n",
      "H_L0.01_weight              0.043\n",
      "MI_dir_L0.01_weight         0.036\n",
      "HH_L0.01_weight             0.035\n",
      "MI_dir_L0.01_variance       0.029\n",
      "HH_L0.1_weight              0.027\n",
      "H_L0.01_mean                0.021\n",
      "MI_dir_L3_weight            0.021\n",
      "MI_dir_L0.01_mean           0.018\n",
      "HH_jit_L0.1_weight          0.016\n",
      "MI_dir_L0.1_mean            0.014\n",
      "MI_dir_L1_weight            0.014\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[18587     0     0]\n",
      " [    0 18646     0]\n",
      " [    0     0 18706]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18587\n",
      "      gafgyt       1.00      1.00      1.00     18646\n",
      "       mirai       1.00      1.00      1.00     18706\n",
      "\n",
      "    accuracy                           1.00     55939\n",
      "   macro avg       1.00      1.00      1.00     55939\n",
      "weighted avg       1.00      1.00      1.00     55939\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(295542, 115), (295542,)\n",
      "                      importance\n",
      "feature                         \n",
      "HH_jit_L5_mean             0.108\n",
      "HH_jit_L0.1_mean           0.096\n",
      "HH_jit_L1_mean             0.078\n",
      "MI_dir_L0.1_weight         0.075\n",
      "HH_jit_L0.01_mean          0.064\n",
      "HH_jit_L3_mean             0.055\n",
      "H_L0.1_weight              0.054\n",
      "HH_jit_L0.01_weight        0.039\n",
      "HH_L0.1_weight             0.039\n",
      "H_L1_weight                0.031\n",
      "MI_dir_L1_weight           0.028\n",
      "MI_dir_L0.01_mean          0.027\n",
      "H_L0.01_variance           0.026\n",
      "H_L0.1_variance            0.025\n",
      "HH_jit_L0.1_weight         0.025\n",
      "MI_dir_L0.01_weight        0.022\n",
      "HH_L0.01_weight            0.020\n",
      "H_L0.01_weight             0.020\n",
      "H_L0.01_mean               0.018\n",
      "MI_dir_L0.1_variance       0.018\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[29478     0     0]\n",
      " [    0 29701     0]\n",
      " [    1     0 29483]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29478\n",
      "      gafgyt       1.00      1.00      1.00     29701\n",
      "       mirai       1.00      1.00      1.00     29484\n",
      "\n",
      "    accuracy                           1.00     88663\n",
      "   macro avg       1.00      1.00      1.00     88663\n",
      "weighted avg       1.00      1.00      1.00     88663\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(104300, 115), (104300,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L3_mean              0.133\n",
      "HH_jit_L0.1_mean            0.070\n",
      "H_L0.01_mean                0.064\n",
      "HH_jit_L0.01_mean           0.053\n",
      "HH_jit_L5_mean              0.051\n",
      "H_L0.1_mean                 0.050\n",
      "MI_dir_L0.01_mean           0.050\n",
      "HpHp_L0.01_weight           0.049\n",
      "HH_jit_L1_mean              0.044\n",
      "HH_L0.01_weight             0.040\n",
      "HpHp_L0.1_weight            0.039\n",
      "HH_L0.1_magnitude           0.037\n",
      "HH_jit_L0.01_weight         0.031\n",
      "HH_jit_L0.1_weight          0.030\n",
      "MI_dir_L0.01_variance       0.029\n",
      "HH_L5_magnitude             0.028\n",
      "HH_L3_magnitude             0.028\n",
      "HpHp_L0.01_magnitude        0.028\n",
      "MI_dir_L0.1_mean            0.021\n",
      "HH_L0.1_weight              0.020\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15642     0]\n",
      " [    0 15648]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15642\n",
      "      gafgyt       1.00      1.00      1.00     15648\n",
      "\n",
      "    accuracy                           1.00     31290\n",
      "   macro avg       1.00      1.00      1.00     31290\n",
      "weighted avg       1.00      1.00      1.00     31290\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(139755, 115), (139755,)\n",
      "                     importance\n",
      "feature                        \n",
      "HH_jit_L1_mean            0.137\n",
      "HH_jit_L0.1_mean          0.083\n",
      "HH_jit_L5_mean            0.082\n",
      "HH_jit_L3_mean            0.074\n",
      "HH_L0.01_magnitude        0.063\n",
      "HH_jit_L0.01_mean         0.063\n",
      "HH_L3_magnitude           0.033\n",
      "HH_L1_magnitude           0.032\n",
      "MI_dir_L0.1_weight        0.030\n",
      "MI_dir_L1_weight          0.028\n",
      "HH_jit_L0.01_weight       0.025\n",
      "HH_L0.1_magnitude         0.025\n",
      "HH_jit_L0.1_weight        0.025\n",
      "MI_dir_L0.1_mean          0.020\n",
      "HpHp_L0.01_weight         0.020\n",
      "HH_L0.01_weight           0.020\n",
      "H_L0.1_weight             0.020\n",
      "HpHp_L1_magnitude         0.017\n",
      "HH_L0.01_mean             0.016\n",
      "H_L0.01_mean              0.013\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[13999     0     0]\n",
      " [    0 13971     0]\n",
      " [    0     0 13957]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     13999\n",
      "      gafgyt       1.00      1.00      1.00     13971\n",
      "       mirai       1.00      1.00      1.00     13957\n",
      "\n",
      "    accuracy                           1.00     41927\n",
      "   macro avg       1.00      1.00      1.00     41927\n",
      "weighted avg       1.00      1.00      1.00     41927\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(58584, 115), (58584,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean              0.119\n",
      "HH_jit_L3_mean              0.115\n",
      "HH_jit_L0.01_mean           0.089\n",
      "HH_jit_L5_mean              0.086\n",
      "HH_jit_L0.1_mean            0.066\n",
      "H_L0.01_mean                0.055\n",
      "HpHp_L0.01_weight           0.035\n",
      "MI_dir_L0.01_variance       0.034\n",
      "H_L0.01_variance            0.033\n",
      "H_L0.1_mean                 0.032\n",
      "MI_dir_L0.1_mean            0.031\n",
      "MI_dir_L0.1_weight          0.028\n",
      "MI_dir_L0.01_mean           0.022\n",
      "H_L0.1_weight               0.021\n",
      "HH_L0.01_weight             0.020\n",
      "H_L5_weight                 0.015\n",
      "MI_dir_L3_weight            0.015\n",
      "H_L1_weight                 0.014\n",
      "HH_L5_magnitude             0.013\n",
      "H_L0.1_variance             0.010\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[5743    0    0]\n",
      " [   0 5859    0]\n",
      " [   0    0 5974]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5743\n",
      "      gafgyt       1.00      1.00      1.00      5859\n",
      "       mirai       1.00      1.00      1.00      5974\n",
      "\n",
      "    accuracy                           1.00     17576\n",
      "   macro avg       1.00      1.00      1.00     17576\n",
      "weighted avg       1.00      1.00      1.00     17576\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with scaling and balanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier_balanced(dataframe[k], k, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def knn_classifier_unbalanced(data, scaling=False):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "    test_scores = []\n",
    "    train_scores = []\n",
    "    error = []\n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        \n",
    "    #neighbors = list(range(1,10))\n",
    "    #for i in neighbors:\n",
    "    knn = KNeighborsClassifier(3)\n",
    "    knn.fit(X_train,y_train)\n",
    "    y_pred = knn.predict(X_test)\n",
    "    print(confusion_matrix(y_test, y_pred))\n",
    "    print(classification_report(y_test, y_pred))\n",
    "    error.append(np.mean(y_pred != y_test))\n",
    "    train_scores.append(knn.score(X_train,y_train))\n",
    "    test_scores.append(knn.score(X_test,y_test))\n",
    "    \n",
    "    print(f'Accuracy on Training set:{train_scores}')\n",
    "    print(f'Accuracy on Test set:{test_scores}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "[[ 14871     33     99]\n",
      " [    93  94724     42]\n",
      " [   108     33 195487]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.99      0.99      0.99     15003\n",
      "      gafgyt       1.00      1.00      1.00     94859\n",
      "       mirai       1.00      1.00      1.00    195628\n",
      "\n",
      "    accuracy                           1.00    305490\n",
      "   macro avg       1.00      1.00      1.00    305490\n",
      "weighted avg       1.00      1.00      1.00    305490\n",
      "\n",
      "Accuracy on Training set:[0.9991835108472408]\n",
      "Accuracy on Test set:[0.9986644407345576]\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "[[  3879     54     83]\n",
      " [    36  92732     38]\n",
      " [    29     23 153889]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.98      0.97      0.97      4016\n",
      "      gafgyt       1.00      1.00      1.00     92806\n",
      "       mirai       1.00      1.00      1.00    153941\n",
      "\n",
      "    accuracy                           1.00    250763\n",
      "   macro avg       0.99      0.99      0.99    250763\n",
      "weighted avg       1.00      1.00      1.00    250763\n",
      "\n",
      "Accuracy on Training set:[0.9994018249466342]\n",
      "Accuracy on Test set:[0.9989512009347471]\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "[[11840    63]\n",
      " [  117 94630]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.99      0.99      0.99     11903\n",
      "      gafgyt       1.00      1.00      1.00     94747\n",
      "\n",
      "    accuracy                           1.00    106650\n",
      "   macro avg       0.99      1.00      1.00    106650\n",
      "weighted avg       1.00      1.00      1.00    106650\n",
      "\n",
      "Accuracy on Training set:[0.9990878038979305]\n",
      "Accuracy on Test set:[0.9983122362869198]\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "[[ 51369     38   1091]\n",
      " [   206  93651     23]\n",
      " [   355      8 182863]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.99      0.98      0.98     52498\n",
      "      gafgyt       1.00      1.00      1.00     93880\n",
      "       mirai       0.99      1.00      1.00    183226\n",
      "\n",
      "    accuracy                           0.99    329604\n",
      "   macro avg       0.99      0.99      0.99    329604\n",
      "weighted avg       0.99      0.99      0.99    329604\n",
      "\n",
      "Accuracy on Training set:[0.995931465543583]\n",
      "Accuracy on Test set:[0.9947785827841895]\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "[[ 18473     40    245]\n",
      " [   104  98395    107]\n",
      " [   278     73 130763]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.98      0.98      0.98     18758\n",
      "      gafgyt       1.00      1.00      1.00     98606\n",
      "       mirai       1.00      1.00      1.00    131114\n",
      "\n",
      "    accuracy                           1.00    248478\n",
      "   macro avg       0.99      0.99      0.99    248478\n",
      "weighted avg       1.00      1.00      1.00    248478\n",
      "\n",
      "Accuracy on Training set:[0.9978543659513404]\n",
      "Accuracy on Test set:[0.9965912475148705]\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "[[ 29083    103    512]\n",
      " [   178  91980     89]\n",
      " [   462     41 128620]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.98      0.98      0.98     29698\n",
      "      gafgyt       1.00      1.00      1.00     92247\n",
      "       mirai       1.00      1.00      1.00    129123\n",
      "\n",
      "    accuracy                           0.99    251068\n",
      "   macro avg       0.99      0.99      0.99    251068\n",
      "weighted avg       0.99      0.99      0.99    251068\n",
      "\n",
      "Accuracy on Training set:[0.9966559865351822]\n",
      "Accuracy on Test set:[0.9944835662051715]\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "[[15592    92]\n",
      " [  219 96664]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.99      0.99      0.99     15684\n",
      "      gafgyt       1.00      1.00      1.00     96883\n",
      "\n",
      "    accuracy                           1.00    112567\n",
      "   macro avg       0.99      1.00      0.99    112567\n",
      "weighted avg       1.00      1.00      1.00    112567\n",
      "\n",
      "Accuracy on Training set:[0.9982905332089623]\n",
      "Accuracy on Test set:[0.9972372009558752]\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "[[ 13849     71    135]\n",
      " [   125  90374     34]\n",
      " [    86     13 154230]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.98      0.99      0.99     14055\n",
      "      gafgyt       1.00      1.00      1.00     90533\n",
      "       mirai       1.00      1.00      1.00    154329\n",
      "\n",
      "    accuracy                           1.00    258917\n",
      "   macro avg       0.99      0.99      0.99    258917\n",
      "weighted avg       1.00      1.00      1.00    258917\n",
      "\n",
      "Accuracy on Training set:[0.9988628444778437]\n",
      "Accuracy on Test set:[0.9982079199125589]\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "[[  5769     20    117]\n",
      " [    66  94412     33]\n",
      " [   147     29 154655]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      benign       0.96      0.98      0.97      5906\n",
      "      gafgyt       1.00      1.00      1.00     94511\n",
      "       mirai       1.00      1.00      1.00    154831\n",
      "\n",
      "    accuracy                           1.00    255248\n",
      "   macro avg       0.99      0.99      0.99    255248\n",
      "weighted avg       1.00      1.00      1.00    255248\n",
      "\n",
      "Accuracy on Training set:[0.9991319356994381]\n",
      "Accuracy on Test set:[0.9983858835328778]\n"
     ]
    }
   ],
   "source": [
    "# without  scaling and unbalanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    knn_classifier_unbalanced(dataframe[k])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n"
     ]
    }
   ],
   "source": [
    "# with  scaling and unbalanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    knn_classifier_unbalanced(dataframe[k], True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def knn_classifier_balanced(data, scaling=False):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "    nm = NearMiss()\n",
    "    \n",
    "    X_res, y_res = nm.fit_resample(X, y)\n",
    "    print(f'Imblearn Shape:{X_res.shape}, {y_res.shape}')\n",
    "    \n",
    "    test_scores = []\n",
    "    train_scores = []\n",
    "    error = []\n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.30, random_state=47)\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        \n",
    "    #neighbors = list(range(1,10))\n",
    "    #for i in neighbors:\n",
    "    knn = KNeighborsClassifier(3)\n",
    "    knn.fit(X_train,y_train)\n",
    "    y_pred = knn.predict(X_test)\n",
    "    print(confusion_matrix(y_test, y_pred))\n",
    "    print(classification_report(y_test, y_pred))\n",
    "    error.append(np.mean(y_pred != y_test))\n",
    "    train_scores.append(knn.score(X_train,y_train))\n",
    "    test_scores.append(knn.score(X_test,y_test))\n",
    "    \n",
    "    print(f'Accuracy on Training set:{train_scores}')\n",
    "    print(f'Accuracy on Test set:{test_scores}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# without  scaling and balanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    knn_classifier_balanced(dataframe[k])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with scaling and balanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    knn_classifier_balanced(dataframe[k], True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Loading data with multi label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_doorbell_multi_df = load_device_data_multi_label(base_directory, file_extension, 'Danmini_Doorbell')\n",
    "ecobee_thermostat_multi_df = load_device_data_multi_label(base_directory, file_extension, 'Ecobee_Thermostat')\n",
    "ennio_doorbell_multi_df = load_device_data_multi_label(base_directory, file_extension, 'Ennio_Doorbell')\n",
    "baby_monitor_multi_df = load_device_data_multi_label(base_directory, file_extension, 'Philips_B120N10_Baby_Monitor')\n",
    "provision_cam1_multi_df = load_device_data_multi_label(base_directory, file_extension, 'Provision_PT_737E_Security_Camera')\n",
    "provision_cam2_multi_df = load_device_data_multi_label(base_directory, file_extension, 'Provision_PT_838_Security_Camera')\n",
    "samsung_cam_multi_df = load_device_data_multi_label(base_directory, file_extension, 'Samsung_SNH_1011_N_Webcam')\n",
    "simplehome_cam1_multi_df = load_device_data_multi_label(base_directory, file_extension, 'SimpleHome_XCS7_1002_WHT_Security_Camera')\n",
    "simplehome_cam2_multi_df = load_device_data_multi_label(base_directory, file_extension, 'SimpleHome_XCS7_1003_WHT_Security_Camera')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataframe_multi = {\"DanminDoorBell_multi \": danmini_doorbell_multi_df, \n",
    "            \"Ecobee_Thermostat_multi\": ecobee_thermostat_multi_df,\n",
    "            \"Ennio_Doorbell_multi\": ennio_doorbell_multi_df,\n",
    "            \"Philips_B120N10_Baby_Monitor_multi\": baby_monitor_multi_df,\n",
    "            \"Provision_PT_737E_Security_Camera_multi\": provision_cam1_multi_df,\n",
    "            \"Provision_PT_838_Security_Camera_multi\": provision_cam2_multi_df,\n",
    "            \"Samsung_SNH_1011_N_Webcam_multi\": samsung_cam_multi_df,\n",
    "            \"SimpleHome_XCS7_1002_WHT_Security_Camera_multi\": simplehome_cam1_multi_df,\n",
    "            \"SimpleHome_XCS7_1003_WHT_Security_Camera_multi\": simplehome_cam2_multi_df\n",
    "            }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_mean                 0.069\n",
      "H_L0.01_mean                0.068\n",
      "MI_dir_L0.1_mean            0.067\n",
      "MI_dir_L0.01_mean           0.057\n",
      "H_L0.1_variance             0.049\n",
      "H_L0.01_variance            0.044\n",
      "H_L1_mean                   0.042\n",
      "MI_dir_L1_mean              0.039\n",
      "MI_dir_L0.1_variance        0.038\n",
      "HH_jit_L1_mean              0.034\n",
      "MI_dir_L0.01_variance       0.033\n",
      "HH_jit_L3_mean              0.028\n",
      "HH_jit_L5_mean              0.025\n",
      "HH_jit_L0.1_mean            0.024\n",
      "HH_jit_L0.01_mean           0.024\n",
      "MI_dir_L0.1_weight          0.023\n",
      "MI_dir_L3_mean              0.022\n",
      "H_L0.1_weight               0.017\n",
      "H_L1_weight                 0.016\n",
      "H_L3_mean                   0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15000     0     0     1     0     1     0     1     0     0     0]\n",
      " [    0 18061     3     0     0     0     0     0     0     0     0]\n",
      " [    0     0  8707     0     0     0     0     0     0     0     0]\n",
      " [    2     0     1  8824     0     0     0     0     0     0     0]\n",
      " [    1     1     0     0 27541    29     0     0     0     0     0]\n",
      " [    6     0     0     0     1 31682     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 30707     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 32314     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 36919     0     0]\n",
      " [    0     0     0     0     2     0     0     0     0 71302     0]\n",
      " [    0     0     0     0     0     0     0     0     0     0 24384]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     15003\n",
      "  gafgyt_combo       1.00      1.00      1.00     18064\n",
      "   gafgyt_junk       1.00      1.00      1.00      8707\n",
      "   gafgyt_scan       1.00      1.00      1.00      8827\n",
      "    gafgyt_tcp       1.00      1.00      1.00     27572\n",
      "    gafgyt_udp       1.00      1.00      1.00     31689\n",
      "     mirai_ack       1.00      1.00      1.00     30707\n",
      "    mirai_scan       1.00      1.00      1.00     32314\n",
      "     mirai_syn       1.00      1.00      1.00     36919\n",
      "     mirai_udp       1.00      1.00      1.00     71304\n",
      "mirai_udpplain       1.00      1.00      1.00     24384\n",
      "\n",
      "      accuracy                           1.00    305490\n",
      "     macro avg       1.00      1.00      1.00    305490\n",
      "  weighted avg       1.00      1.00      1.00    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.086\n",
      "H_L0.01_mean                0.074\n",
      "MI_dir_L0.01_variance       0.055\n",
      "H_L0.01_variance            0.054\n",
      "MI_dir_L0.1_mean            0.043\n",
      "HH_jit_L0.01_mean           0.038\n",
      "H_L0.1_mean                 0.037\n",
      "HH_jit_L0.1_mean            0.036\n",
      "HH_jit_L1_mean              0.033\n",
      "HH_jit_L3_mean              0.032\n",
      "HH_jit_L5_mean              0.030\n",
      "H_L0.1_variance             0.027\n",
      "H_L0.1_weight               0.024\n",
      "H_L1_weight                 0.021\n",
      "H_L1_mean                   0.021\n",
      "H_L3_mean                   0.020\n",
      "MI_dir_L0.1_variance        0.019\n",
      "MI_dir_L0.1_weight          0.018\n",
      "MI_dir_L3_mean              0.018\n",
      "MI_dir_L1_mean              0.018\n",
      "Random Forest: Accuracy=0.998\n",
      "Random Forest: f1-score=0.998\n",
      "Random Forest: Confusion Matrix [[ 4015     0     0     1     0     0     0     0     0     0     0]\n",
      " [    0 15937     4     1     0     0     0     0     0     0     0]\n",
      " [    0     3  9082     0     1     1     0     0     0     0     0]\n",
      " [    0     0     0  8286     0     0     0     0     0     0     0]\n",
      " [    0     1     0     1 27829   492     0     0     0     0     0]\n",
      " [    0     0     0     0     0 31168     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 34041     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 13056     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35094     0     0]\n",
      " [    0     0     0     0     0     0     0     1     0 45554     0]\n",
      " [    1     0     0     0     0     0     0     1     0     4 26188]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      4016\n",
      "  gafgyt_combo       1.00      1.00      1.00     15942\n",
      "   gafgyt_junk       1.00      1.00      1.00      9087\n",
      "   gafgyt_scan       1.00      1.00      1.00      8286\n",
      "    gafgyt_tcp       1.00      0.98      0.99     28323\n",
      "    gafgyt_udp       0.98      1.00      0.99     31168\n",
      "     mirai_ack       1.00      1.00      1.00     34041\n",
      "    mirai_scan       1.00      1.00      1.00     13057\n",
      "     mirai_syn       1.00      1.00      1.00     35094\n",
      "     mirai_udp       1.00      1.00      1.00     45555\n",
      "mirai_udpplain       1.00      1.00      1.00     26194\n",
      "\n",
      "      accuracy                           1.00    250763\n",
      "     macro avg       1.00      1.00      1.00    250763\n",
      "  weighted avg       1.00      1.00      1.00    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.127\n",
      "HH_jit_L1_mean              0.077\n",
      "HH_jit_L5_mean              0.076\n",
      "HH_jit_L0.01_mean           0.075\n",
      "HH_jit_L3_mean              0.061\n",
      "H_L0.01_mean                0.035\n",
      "HH_L0.01_magnitude          0.032\n",
      "H_L0.01_weight              0.030\n",
      "MI_dir_L0.1_mean            0.029\n",
      "MI_dir_L0.01_mean           0.029\n",
      "MI_dir_L0.1_weight          0.028\n",
      "HH_L0.01_weight             0.025\n",
      "MI_dir_L0.01_weight         0.022\n",
      "H_L0.1_weight               0.021\n",
      "HH_jit_L0.1_weight          0.018\n",
      "H_L0.1_mean                 0.018\n",
      "H_L0.01_variance            0.017\n",
      "MI_dir_L0.01_variance       0.017\n",
      "HH_L0.1_weight              0.016\n",
      "H_L1_weight                 0.016\n",
      "Random Forest: Accuracy=0.992\n",
      "Random Forest: f1-score=0.992\n",
      "Random Forest: Confusion Matrix [[11903     0     0     0     0     0]\n",
      " [    0 16021     1     0     2     0]\n",
      " [    0     0  8937     0     1     0]\n",
      " [    0     0     0  8519     1     0]\n",
      " [    0     0     0     0 29475   799]\n",
      " [    0     0     0     0     3 30988]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "gafgyt_combo       1.00      1.00      1.00     16024\n",
      " gafgyt_junk       1.00      1.00      1.00      8938\n",
      " gafgyt_scan       1.00      1.00      1.00      8520\n",
      "  gafgyt_tcp       1.00      0.97      0.99     30274\n",
      "  gafgyt_udp       0.97      1.00      0.99     30991\n",
      "\n",
      "    accuracy                           0.99    106650\n",
      "   macro avg       1.00      1.00      1.00    106650\n",
      "weighted avg       0.99      0.99      0.99    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.065\n",
      "MI_dir_L0.01_mean           0.059\n",
      "MI_dir_L0.1_mean            0.056\n",
      "H_L0.1_mean                 0.051\n",
      "HH_jit_L0.01_mean           0.039\n",
      "MI_dir_L0.01_variance       0.037\n",
      "H_L0.01_variance            0.031\n",
      "H_L1_mean                   0.028\n",
      "HH_jit_L0.1_mean            0.028\n",
      "HH_jit_L5_mean              0.027\n",
      "MI_dir_L1_mean              0.026\n",
      "H_L0.1_variance             0.025\n",
      "MI_dir_L0.1_variance        0.024\n",
      "H_L3_mean                   0.021\n",
      "HH_jit_L1_mean              0.021\n",
      "HpHp_L0.01_radius           0.020\n",
      "H_L0.1_weight               0.019\n",
      "HH_jit_L3_mean              0.019\n",
      "H_L1_weight                 0.017\n",
      "H_L5_mean                   0.016\n",
      "Random Forest: Accuracy=0.997\n",
      "Random Forest: f1-score=0.997\n",
      "Random Forest: Confusion Matrix [[52497     0     0     0     0     1     0     0     0     0     0]\n",
      " [    0 17516     1     1     0     0     0     0     0     0     0]\n",
      " [    1     2  8473     0     0     1     0     0     0     0     0]\n",
      " [    0     0     0  8229     1     0     0     0     0     0     0]\n",
      " [    0     0     0     0 26904  1001     0     0     0     0     0]\n",
      " [    8     0     0     0     0 31742     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 27317     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 31097     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35288     0     0]\n",
      " [    0     0     0     0     1     0     0     0     0 65060     0]\n",
      " [    0     0     0     0     0     0     0     0     0     0 24463]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     52498\n",
      "  gafgyt_combo       1.00      1.00      1.00     17518\n",
      "   gafgyt_junk       1.00      1.00      1.00      8477\n",
      "   gafgyt_scan       1.00      1.00      1.00      8230\n",
      "    gafgyt_tcp       1.00      0.96      0.98     27905\n",
      "    gafgyt_udp       0.97      1.00      0.98     31750\n",
      "     mirai_ack       1.00      1.00      1.00     27317\n",
      "    mirai_scan       1.00      1.00      1.00     31097\n",
      "     mirai_syn       1.00      1.00      1.00     35288\n",
      "     mirai_udp       1.00      1.00      1.00     65061\n",
      "mirai_udpplain       1.00      1.00      1.00     24463\n",
      "\n",
      "      accuracy                           1.00    329604\n",
      "     macro avg       1.00      1.00      1.00    329604\n",
      "  weighted avg       1.00      1.00      1.00    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.067\n",
      "MI_dir_L0.01_mean           0.067\n",
      "MI_dir_L0.01_variance       0.055\n",
      "H_L0.01_variance            0.049\n",
      "HH_jit_L0.1_mean            0.047\n",
      "H_L0.1_variance             0.047\n",
      "H_L0.1_mean                 0.047\n",
      "MI_dir_L0.1_mean            0.045\n",
      "HH_jit_L3_mean              0.035\n",
      "HH_jit_L1_mean              0.034\n",
      "MI_dir_L0.1_variance        0.034\n",
      "HH_jit_L0.01_mean           0.031\n",
      "H_L0.1_weight               0.030\n",
      "MI_dir_L0.1_weight          0.029\n",
      "H_L0.01_weight              0.028\n",
      "HH_jit_L5_mean              0.027\n",
      "MI_dir_L0.01_weight         0.023\n",
      "H_L3_mean                   0.022\n",
      "H_L1_mean                   0.021\n",
      "MI_dir_L1_mean              0.017\n",
      "Random Forest: Accuracy=0.993\n",
      "Random Forest: f1-score=0.993\n",
      "Random Forest: Confusion Matrix [[18758     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 18404     2     0     0     0     0     0     0     0     0]\n",
      " [    0     2  9174     0     1     0     0     0     0     0     0]\n",
      " [    1     0     0  8619     1     0     0     0     0     0     0]\n",
      " [    1     0     0     0 29330  1808     0     0     0     0     0]\n",
      " [    3     0     0     0     2 31258     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 18165     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 28964     0     0     0]\n",
      " [    0     0     0     0     0     0     2     0 19841     0     0]\n",
      " [    0     0     0     0     0     0     0     0     0 46981     0]\n",
      " [    0     0     0     0     0     0     0     0     0     0 17160]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     18758\n",
      "  gafgyt_combo       1.00      1.00      1.00     18406\n",
      "   gafgyt_junk       1.00      1.00      1.00      9177\n",
      "   gafgyt_scan       1.00      1.00      1.00      8621\n",
      "    gafgyt_tcp       1.00      0.94      0.97     31139\n",
      "    gafgyt_udp       0.95      1.00      0.97     31263\n",
      "     mirai_ack       1.00      1.00      1.00     18165\n",
      "    mirai_scan       1.00      1.00      1.00     28965\n",
      "     mirai_syn       1.00      1.00      1.00     19843\n",
      "     mirai_udp       1.00      1.00      1.00     46981\n",
      "mirai_udpplain       1.00      1.00      1.00     17160\n",
      "\n",
      "      accuracy                           0.99    248478\n",
      "     macro avg       0.99      0.99      0.99    248478\n",
      "  weighted avg       0.99      0.99      0.99    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.067\n",
      "MI_dir_L0.01_mean           0.061\n",
      "MI_dir_L0.1_mean            0.059\n",
      "H_L0.1_mean                 0.059\n",
      "H_L0.01_variance            0.051\n",
      "HH_jit_L0.01_mean           0.049\n",
      "MI_dir_L0.01_variance       0.046\n",
      "MI_dir_L0.1_variance        0.044\n",
      "H_L0.1_variance             0.042\n",
      "HH_jit_L0.1_mean            0.036\n",
      "MI_dir_L0.1_weight          0.033\n",
      "H_L0.01_weight              0.032\n",
      "HH_jit_L3_mean              0.030\n",
      "H_L0.1_weight               0.030\n",
      "HH_jit_L1_mean              0.029\n",
      "HH_jit_L5_mean              0.027\n",
      "MI_dir_L1_mean              0.025\n",
      "MI_dir_L0.01_weight         0.023\n",
      "H_L1_mean                   0.019\n",
      "MI_dir_L3_mean              0.018\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[29698     0     0     0     0     0     0     0     0     0     0]\n",
      " [    2 17163     2     1     1     0     0     0     0     0     0]\n",
      " [    0     4  8587     1     0     0     0     0     0     0     0]\n",
      " [    0     0     0  8561     2     0     0     0     0     0     0]\n",
      " [    0     0     0     2 26571     5     0     0     0     0     0]\n",
      " [    2     0     0     1     0 31342     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 17476     0     1     0     0]\n",
      " [    0     0     0     0     0     0     0 29206     0     0     0]\n",
      " [    0     0     0     0     0     0     0     1 18635     0     0]\n",
      " [    0     0     0     0     0     0     0     0     0 47548     0]\n",
      " [    0     0     0     0     0     0     0     0     0     1 16255]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     29698\n",
      "  gafgyt_combo       1.00      1.00      1.00     17169\n",
      "   gafgyt_junk       1.00      1.00      1.00      8592\n",
      "   gafgyt_scan       1.00      1.00      1.00      8563\n",
      "    gafgyt_tcp       1.00      1.00      1.00     26578\n",
      "    gafgyt_udp       1.00      1.00      1.00     31345\n",
      "     mirai_ack       1.00      1.00      1.00     17477\n",
      "    mirai_scan       1.00      1.00      1.00     29206\n",
      "     mirai_syn       1.00      1.00      1.00     18636\n",
      "     mirai_udp       1.00      1.00      1.00     47548\n",
      "mirai_udpplain       1.00      1.00      1.00     16256\n",
      "\n",
      "      accuracy                           1.00    251068\n",
      "     macro avg       1.00      1.00      1.00    251068\n",
      "  weighted avg       1.00      1.00      1.00    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.095\n",
      "HH_jit_L3_mean              0.081\n",
      "HH_jit_L5_mean              0.078\n",
      "HH_jit_L0.01_mean           0.070\n",
      "HH_jit_L1_mean              0.068\n",
      "MI_dir_L0.1_mean            0.053\n",
      "MI_dir_L0.01_mean           0.043\n",
      "H_L0.01_mean                0.034\n",
      "H_L0.1_mean                 0.031\n",
      "MI_dir_L0.01_weight         0.027\n",
      "H_L1_mean                   0.023\n",
      "HH_L0.01_magnitude          0.022\n",
      "MI_dir_L0.1_weight          0.022\n",
      "H_L0.01_weight              0.019\n",
      "MI_dir_L3_mean              0.018\n",
      "H_L0.1_weight               0.017\n",
      "MI_dir_L0.01_variance       0.015\n",
      "HH_L0.1_weight              0.015\n",
      "HH_jit_L0.01_weight         0.013\n",
      "HH_L5_magnitude             0.013\n",
      "Random Forest: Accuracy=0.999\n",
      "Random Forest: f1-score=0.999\n",
      "Random Forest: Confusion Matrix [[15683     0     0     0     0     1]\n",
      " [    0 17688     1     0     1     0]\n",
      " [    0     0  8606     0     0     0]\n",
      " [    0     0     0  8195     1     1]\n",
      " [    3     0     0     0 29254    80]\n",
      " [    3     0     0     0     1 33049]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "gafgyt_combo       1.00      1.00      1.00     17690\n",
      " gafgyt_junk       1.00      1.00      1.00      8606\n",
      " gafgyt_scan       1.00      1.00      1.00      8197\n",
      "  gafgyt_tcp       1.00      1.00      1.00     29337\n",
      "  gafgyt_udp       1.00      1.00      1.00     33053\n",
      "\n",
      "    accuracy                           1.00    112567\n",
      "   macro avg       1.00      1.00      1.00    112567\n",
      "weighted avg       1.00      1.00      1.00    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.084\n",
      "H_L0.01_mean                0.076\n",
      "H_L0.01_variance            0.050\n",
      "MI_dir_L0.01_variance       0.050\n",
      "H_L0.1_mean                 0.040\n",
      "MI_dir_L0.1_mean            0.039\n",
      "HH_jit_L0.01_mean           0.035\n",
      "HH_jit_L3_mean              0.033\n",
      "HH_jit_L0.1_mean            0.033\n",
      "HH_jit_L1_mean              0.032\n",
      "MI_dir_L0.1_weight          0.029\n",
      "H_L0.1_variance             0.027\n",
      "HH_jit_L5_mean              0.027\n",
      "MI_dir_L0.1_variance        0.025\n",
      "H_L0.1_weight               0.025\n",
      "MI_dir_L1_mean              0.023\n",
      "H_L0.01_weight              0.023\n",
      "H_L1_mean                   0.023\n",
      "MI_dir_L0.01_weight         0.020\n",
      "MI_dir_L1_variance          0.018\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[14055     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 16314     2     2     0     0     0     0     0     0     0]\n",
      " [    0     2  8620     1     0     0     0     0     0     0     0]\n",
      " [    0     0     1  8210     1     1     0     0     0     0     0]\n",
      " [    0     0     1     1 26446     2     0     0     0     0     0]\n",
      " [    4     0     1     1     1 30922     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 33471     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 13876     0     0     1]\n",
      " [    0     0     0     0     0     0     0     0 37909     0     0]\n",
      " [    0     0     0     0     0     0     0     0     0 45602     1]\n",
      " [    0     0     0     0     0     0     0     0     0     1 23467]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     14055\n",
      "  gafgyt_combo       1.00      1.00      1.00     16318\n",
      "   gafgyt_junk       1.00      1.00      1.00      8623\n",
      "   gafgyt_scan       1.00      1.00      1.00      8213\n",
      "    gafgyt_tcp       1.00      1.00      1.00     26450\n",
      "    gafgyt_udp       1.00      1.00      1.00     30929\n",
      "     mirai_ack       1.00      1.00      1.00     33471\n",
      "    mirai_scan       1.00      1.00      1.00     13878\n",
      "     mirai_syn       1.00      1.00      1.00     37909\n",
      "     mirai_udp       1.00      1.00      1.00     45603\n",
      "mirai_udpplain       1.00      1.00      1.00     23468\n",
      "\n",
      "      accuracy                           1.00    258917\n",
      "     macro avg       1.00      1.00      1.00    258917\n",
      "  weighted avg       1.00      1.00      1.00    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.085\n",
      "MI_dir_L0.01_mean           0.078\n",
      "MI_dir_L0.01_variance       0.056\n",
      "H_L0.01_variance            0.039\n",
      "HH_jit_L1_mean              0.038\n",
      "H_L0.1_mean                 0.037\n",
      "MI_dir_L0.1_mean            0.035\n",
      "HH_jit_L3_mean              0.034\n",
      "HH_jit_L0.1_mean            0.032\n",
      "H_L0.1_variance             0.030\n",
      "MI_dir_L0.1_weight          0.030\n",
      "HH_jit_L5_mean              0.029\n",
      "HH_jit_L0.01_mean           0.026\n",
      "H_L1_mean                   0.025\n",
      "H_L0.1_weight               0.022\n",
      "H_L0.01_weight              0.021\n",
      "MI_dir_L1_variance          0.017\n",
      "HH_L0.1_magnitude           0.017\n",
      "MI_dir_L1_mean              0.016\n",
      "H_L1_variance               0.016\n",
      "Random Forest: Accuracy=0.993\n",
      "Random Forest: f1-score=0.993\n",
      "Random Forest: Confusion Matrix [[ 5906     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 17831     0     0     1     0     0     0     0     0     0]\n",
      " [    0     6  8259     0     0     1     0     0     0     0     0]\n",
      " [    0     2     0  8568     0     0     0     0     0     0     0]\n",
      " [    0     0     0     0 27332  1836     0     0     0     0     0]\n",
      " [    0     0     0     0     0 30675     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 32254     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 13086     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 36878     1     0]\n",
      " [    0     0     0     0     0     0     0     0     0 47134     0]\n",
      " [    2     0     0     0     0     0     2     0     0     0 25473]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      5906\n",
      "  gafgyt_combo       1.00      1.00      1.00     17832\n",
      "   gafgyt_junk       1.00      1.00      1.00      8266\n",
      "   gafgyt_scan       1.00      1.00      1.00      8570\n",
      "    gafgyt_tcp       1.00      0.94      0.97     29168\n",
      "    gafgyt_udp       0.94      1.00      0.97     30675\n",
      "     mirai_ack       1.00      1.00      1.00     32254\n",
      "    mirai_scan       1.00      1.00      1.00     13087\n",
      "     mirai_syn       1.00      1.00      1.00     36879\n",
      "     mirai_udp       1.00      1.00      1.00     47134\n",
      "mirai_udpplain       1.00      1.00      1.00     25477\n",
      "\n",
      "      accuracy                           0.99    255248\n",
      "     macro avg       0.99      0.99      0.99    255248\n",
      "  weighted avg       0.99      0.99      0.99    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with out scaling unbalanced data\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier(dataframe_multi[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.089\n",
      "MI_dir_L0.01_mean           0.072\n",
      "H_L0.01_variance            0.069\n",
      "H_L0.1_mean                 0.066\n",
      "MI_dir_L0.01_variance       0.063\n",
      "H_L0.1_variance             0.052\n",
      "MI_dir_L0.1_mean            0.047\n",
      "MI_dir_L1_mean              0.036\n",
      "H_L3_mean                   0.030\n",
      "H_L1_mean                   0.029\n",
      "MI_dir_L0.1_variance        0.029\n",
      "MI_dir_L0.01_weight         0.024\n",
      "H_L0.1_weight               0.023\n",
      "MI_dir_L1_weight            0.021\n",
      "MI_dir_L0.1_weight          0.019\n",
      "H_L0.01_weight              0.018\n",
      "H_L1_weight                 0.017\n",
      "HH_L0.1_magnitude           0.014\n",
      "H_L5_mean                   0.013\n",
      "MI_dir_L3_mean              0.013\n",
      "Random Forest: Accuracy=0.910\n",
      "Random Forest: f1-score=0.878\n",
      "Random Forest: Confusion Matrix [[15002     0     0     0     0     0     0     0     1     0     0]\n",
      " [    0 18062     2     0     0     0     0     0     0     0     0]\n",
      " [    0     0  8707     0     0     0     0     0     0     0     0]\n",
      " [    1     0     1  8824     1     0     0     0     0     0     0]\n",
      " [    1     1     0     0    31 27539     0     0     0     0     0]\n",
      " [    6     0     0     0     1 31682     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 30707     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 32314     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 36919     0     0]\n",
      " [    0     0     0     0     0     1     0     0     0 71303     0]\n",
      " [    0     0     0     0     0     0     1     0     0     0 24383]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     15003\n",
      "  gafgyt_combo       1.00      1.00      1.00     18064\n",
      "   gafgyt_junk       1.00      1.00      1.00      8707\n",
      "   gafgyt_scan       1.00      1.00      1.00      8827\n",
      "    gafgyt_tcp       0.94      0.00      0.00     27572\n",
      "    gafgyt_udp       0.53      1.00      0.70     31689\n",
      "     mirai_ack       1.00      1.00      1.00     30707\n",
      "    mirai_scan       1.00      1.00      1.00     32314\n",
      "     mirai_syn       1.00      1.00      1.00     36919\n",
      "     mirai_udp       1.00      1.00      1.00     71304\n",
      "mirai_udpplain       1.00      1.00      1.00     24384\n",
      "\n",
      "      accuracy                           0.91    305490\n",
      "     macro avg       0.95      0.91      0.88    305490\n",
      "  weighted avg       0.95      0.91      0.88    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.090\n",
      "H_L0.01_mean                0.077\n",
      "MI_dir_L0.01_variance       0.057\n",
      "H_L0.01_variance            0.057\n",
      "MI_dir_L0.1_mean            0.048\n",
      "H_L0.1_mean                 0.046\n",
      "MI_dir_L0.1_variance        0.032\n",
      "H_L0.1_variance             0.028\n",
      "MI_dir_L0.1_weight          0.027\n",
      "MI_dir_L5_mean              0.023\n",
      "H_L1_mean                   0.023\n",
      "MI_dir_L1_weight            0.022\n",
      "H_L0.1_weight               0.021\n",
      "H_L3_mean                   0.019\n",
      "MI_dir_L1_mean              0.019\n",
      "MI_dir_L0.01_weight         0.018\n",
      "H_L1_weight                 0.016\n",
      "H_L0.01_weight              0.015\n",
      "H_L1_variance               0.014\n",
      "H_L5_mean                   0.014\n",
      "Random Forest: Accuracy=0.887\n",
      "Random Forest: f1-score=0.848\n",
      "Random Forest: Confusion Matrix [[ 4016     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 15937     3     2     0     0     0     0     0     0     0]\n",
      " [    0     4  9081     0     1     1     0     0     0     0     0]\n",
      " [    0     0     0  8286     0     0     0     0     0     0     0]\n",
      " [    0     1     0     1    21 28300     0     0     0     0     0]\n",
      " [    0     0     0     0     0 31168     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 34041     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 13056     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35094     0     0]\n",
      " [    0     0     0     0     0     0     0     1     0 45554     0]\n",
      " [    0     0     0     0     0     0     0     1     0     4 26189]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      4016\n",
      "  gafgyt_combo       1.00      1.00      1.00     15942\n",
      "   gafgyt_junk       1.00      1.00      1.00      9087\n",
      "   gafgyt_scan       1.00      1.00      1.00      8286\n",
      "    gafgyt_tcp       0.95      0.00      0.00     28323\n",
      "    gafgyt_udp       0.52      1.00      0.69     31168\n",
      "     mirai_ack       1.00      1.00      1.00     34041\n",
      "    mirai_scan       1.00      1.00      1.00     13057\n",
      "     mirai_syn       1.00      1.00      1.00     35094\n",
      "     mirai_udp       1.00      1.00      1.00     45555\n",
      "mirai_udpplain       1.00      1.00      1.00     26194\n",
      "\n",
      "      accuracy                           0.89    250763\n",
      "     macro avg       0.95      0.91      0.88    250763\n",
      "  weighted avg       0.94      0.89      0.85    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.065\n",
      "H_L0.01_mean                0.055\n",
      "MI_dir_L0.01_weight         0.054\n",
      "H_L0.01_weight              0.049\n",
      "H_L0.1_weight               0.040\n",
      "H_L0.1_mean                 0.039\n",
      "HH_L1_magnitude             0.038\n",
      "H_L0.01_variance            0.032\n",
      "MI_dir_L0.01_variance       0.030\n",
      "HH_L0.01_magnitude          0.028\n",
      "HH_L0.1_weight              0.027\n",
      "HH_jit_L0.01_weight         0.026\n",
      "HH_L0.1_magnitude           0.026\n",
      "HH_jit_L0.1_mean            0.024\n",
      "MI_dir_L0.1_weight          0.023\n",
      "HH_jit_L0.01_mean           0.023\n",
      "HH_jit_L1_mean              0.023\n",
      "HH_L3_magnitude             0.021\n",
      "HH_jit_L0.1_weight          0.020\n",
      "HH_L1_weight                0.020\n",
      "Random Forest: Accuracy=0.716\n",
      "Random Forest: f1-score=0.621\n",
      "Random Forest: Confusion Matrix [[11903     0     0     0     0     0]\n",
      " [    0 16021     1     0     2     0]\n",
      " [    0     0  8937     0     1     0]\n",
      " [    0     1     0  8518     1     0]\n",
      " [    0     1     0     0    25 30248]\n",
      " [    1     0     0     0     3 30987]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "gafgyt_combo       1.00      1.00      1.00     16024\n",
      " gafgyt_junk       1.00      1.00      1.00      8938\n",
      " gafgyt_scan       1.00      1.00      1.00      8520\n",
      "  gafgyt_tcp       0.78      0.00      0.00     30274\n",
      "  gafgyt_udp       0.51      1.00      0.67     30991\n",
      "\n",
      "    accuracy                           0.72    106650\n",
      "   macro avg       0.88      0.83      0.78    106650\n",
      "weighted avg       0.79      0.72      0.62    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.077\n",
      "H_L0.01_mean                0.069\n",
      "MI_dir_L0.1_mean            0.063\n",
      "H_L0.1_mean                 0.060\n",
      "MI_dir_L0.01_variance       0.049\n",
      "H_L0.01_variance            0.036\n",
      "H_L0.1_variance             0.033\n",
      "MI_dir_L0.1_variance        0.032\n",
      "H_L1_weight                 0.031\n",
      "H_L1_mean                   0.029\n",
      "H_L0.1_weight               0.028\n",
      "HH_L0.01_std                0.025\n",
      "HpHp_L0.01_std              0.024\n",
      "MI_dir_L1_mean              0.019\n",
      "MI_dir_L3_mean              0.019\n",
      "H_L5_mean                   0.018\n",
      "HpHp_L0.1_radius            0.017\n",
      "HH_L0.01_radius             0.016\n",
      "MI_dir_L0.1_weight          0.016\n",
      "HpHp_L0.01_radius           0.016\n",
      "Random Forest: Accuracy=0.915\n",
      "Random Forest: f1-score=0.886\n",
      "Random Forest: Confusion Matrix [[52498     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 17516     1     1     0     0     0     0     0     0     0]\n",
      " [    1     2  8473     0     0     1     0     0     0     0     0]\n",
      " [    0     0     0  8229     1     0     0     0     0     0     0]\n",
      " [    0     0     0     0    26 27879     0     0     0     0     0]\n",
      " [    8     0     0     0     1 31741     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 27317     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 31097     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35288     0     0]\n",
      " [    0     0     0     0     0     1     0     0     0 65060     0]\n",
      " [    0     0     0     0     0     0     0     0     0     0 24463]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     52498\n",
      "  gafgyt_combo       1.00      1.00      1.00     17518\n",
      "   gafgyt_junk       1.00      1.00      1.00      8477\n",
      "   gafgyt_scan       1.00      1.00      1.00      8230\n",
      "    gafgyt_tcp       0.93      0.00      0.00     27905\n",
      "    gafgyt_udp       0.53      1.00      0.69     31750\n",
      "     mirai_ack       1.00      1.00      1.00     27317\n",
      "    mirai_scan       1.00      1.00      1.00     31097\n",
      "     mirai_syn       1.00      1.00      1.00     35288\n",
      "     mirai_udp       1.00      1.00      1.00     65061\n",
      "mirai_udpplain       1.00      1.00      1.00     24463\n",
      "\n",
      "      accuracy                           0.92    329604\n",
      "     macro avg       0.95      0.91      0.88    329604\n",
      "  weighted avg       0.95      0.92      0.89    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.089\n",
      "MI_dir_L0.01_mean           0.072\n",
      "H_L0.01_variance            0.065\n",
      "MI_dir_L0.01_variance       0.058\n",
      "H_L0.1_mean                 0.057\n",
      "H_L0.1_variance             0.055\n",
      "MI_dir_L0.1_variance        0.053\n",
      "MI_dir_L0.1_mean            0.043\n",
      "MI_dir_L0.01_weight         0.040\n",
      "H_L0.1_weight               0.032\n",
      "H_L0.01_weight              0.030\n",
      "MI_dir_L0.1_weight          0.029\n",
      "H_L1_mean                   0.026\n",
      "H_L3_mean                   0.018\n",
      "HH_jit_L0.01_mean           0.017\n",
      "H_L1_variance               0.016\n",
      "HH_jit_L0.1_mean            0.015\n",
      "MI_dir_L1_mean              0.013\n",
      "H_L1_weight                 0.011\n",
      "HH_jit_L0.01_weight         0.011\n",
      "Random Forest: Accuracy=0.874\n",
      "Random Forest: f1-score=0.833\n",
      "Random Forest: Confusion Matrix [[18758     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 18406     0     0     0     0     0     0     0     0     0]\n",
      " [    0     2  9174     0     1     0     0     0     0     0     0]\n",
      " [    1     0     0  8619     1     0     0     0     0     0     0]\n",
      " [    1     0     0     0 31138     0     0     0     0     0     0]\n",
      " [    3     0     0     0 31232    28     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 18165     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 28964     0     0     0]\n",
      " [    0     0     0     0     0     0     2     0 19841     0     0]\n",
      " [    0     0     0     0     0     0     0     0     0 46981     0]\n",
      " [    0     0     0     0     0     0     0     0     0     0 17160]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     18758\n",
      "  gafgyt_combo       1.00      1.00      1.00     18406\n",
      "   gafgyt_junk       1.00      1.00      1.00      9177\n",
      "   gafgyt_scan       1.00      1.00      1.00      8621\n",
      "    gafgyt_tcp       0.50      1.00      0.67     31139\n",
      "    gafgyt_udp       1.00      0.00      0.00     31263\n",
      "     mirai_ack       1.00      1.00      1.00     18165\n",
      "    mirai_scan       1.00      1.00      1.00     28965\n",
      "     mirai_syn       1.00      1.00      1.00     19843\n",
      "     mirai_udp       1.00      1.00      1.00     46981\n",
      "mirai_udpplain       1.00      1.00      1.00     17160\n",
      "\n",
      "      accuracy                           0.87    248478\n",
      "     macro avg       0.95      0.91      0.88    248478\n",
      "  weighted avg       0.94      0.87      0.83    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.091\n",
      "MI_dir_L0.01_mean           0.080\n",
      "MI_dir_L0.01_variance       0.060\n",
      "H_L0.1_mean                 0.056\n",
      "H_L0.1_variance             0.053\n",
      "H_L0.01_variance            0.052\n",
      "MI_dir_L0.1_mean            0.047\n",
      "MI_dir_L0.1_weight          0.044\n",
      "MI_dir_L0.1_variance        0.041\n",
      "H_L1_mean                   0.038\n",
      "H_L0.01_weight              0.038\n",
      "H_L0.1_weight               0.030\n",
      "MI_dir_L0.01_weight         0.029\n",
      "HH_jit_L0.01_mean           0.023\n",
      "MI_dir_L1_mean              0.018\n",
      "MI_dir_L1_weight            0.016\n",
      "MI_dir_L5_mean              0.014\n",
      "HH_jit_L1_mean              0.013\n",
      "MI_dir_L3_mean              0.012\n",
      "H_L5_mean                   0.011\n",
      "Random Forest: Accuracy=0.894\n",
      "Random Forest: f1-score=0.857\n",
      "Random Forest: Confusion Matrix [[29696     0     0     0     0     0     2     0     0     0     0]\n",
      " [    1 17163     3     1     1     0     0     0     0     0     0]\n",
      " [    0     3  8588     1     0     0     0     0     0     0     0]\n",
      " [    0     0     0  8561     2     0     0     0     0     0     0]\n",
      " [    1     0     0     1    18 26558     0     0     0     0     0]\n",
      " [    2     0     0     1     0 31342     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 17476     0     1     0     0]\n",
      " [    0     0     0     0     0     0     0 29206     0     0     0]\n",
      " [    0     0     0     0     0     0     0     1 18635     0     0]\n",
      " [    0     0     0     0     0     0     0     0     0 47548     0]\n",
      " [    0     0     0     0     0     0     0     0     0     1 16255]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     29698\n",
      "  gafgyt_combo       1.00      1.00      1.00     17169\n",
      "   gafgyt_junk       1.00      1.00      1.00      8592\n",
      "   gafgyt_scan       1.00      1.00      1.00      8563\n",
      "    gafgyt_tcp       0.86      0.00      0.00     26578\n",
      "    gafgyt_udp       0.54      1.00      0.70     31345\n",
      "     mirai_ack       1.00      1.00      1.00     17477\n",
      "    mirai_scan       1.00      1.00      1.00     29206\n",
      "     mirai_syn       1.00      1.00      1.00     18636\n",
      "     mirai_udp       1.00      1.00      1.00     47548\n",
      "mirai_udpplain       1.00      1.00      1.00     16256\n",
      "\n",
      "      accuracy                           0.89    251068\n",
      "     macro avg       0.95      0.91      0.88    251068\n",
      "  weighted avg       0.93      0.89      0.86    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.077\n",
      "MI_dir_L0.1_mean            0.072\n",
      "MI_dir_L0.01_mean           0.065\n",
      "H_L0.1_mean                 0.054\n",
      "MI_dir_L1_mean              0.041\n",
      "H_L0.01_weight              0.040\n",
      "MI_dir_L0.01_weight         0.040\n",
      "H_L1_mean                   0.037\n",
      "H_L0.1_weight               0.032\n",
      "MI_dir_L0.1_weight          0.028\n",
      "H_L0.01_variance            0.026\n",
      "HH_jit_L0.01_mean           0.024\n",
      "MI_dir_L0.01_variance       0.022\n",
      "HH_jit_L3_mean              0.022\n",
      "HH_jit_L0.01_weight         0.020\n",
      "HH_L0.01_weight             0.019\n",
      "HpHp_L0.01_magnitude        0.019\n",
      "HH_jit_L5_mean              0.019\n",
      "HH_jit_L1_mean              0.019\n",
      "HH_jit_L1_weight            0.018\n",
      "Random Forest: Accuracy=0.740\n",
      "Random Forest: f1-score=0.650\n",
      "Random Forest: Confusion Matrix [[15684     0     0     0     0     0]\n",
      " [    0 17687     2     0     1     0]\n",
      " [    0     0  8606     0     0     0]\n",
      " [    0     0     0  8195     1     1]\n",
      " [    4     0     0     0    25 29308]\n",
      " [    3     0     0     0     1 33049]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "gafgyt_combo       1.00      1.00      1.00     17690\n",
      " gafgyt_junk       1.00      1.00      1.00      8606\n",
      " gafgyt_scan       1.00      1.00      1.00      8197\n",
      "  gafgyt_tcp       0.89      0.00      0.00     29337\n",
      "  gafgyt_udp       0.53      1.00      0.69     33053\n",
      "\n",
      "    accuracy                           0.74    112567\n",
      "   macro avg       0.90      0.83      0.78    112567\n",
      "weighted avg       0.83      0.74      0.65    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.102\n",
      "MI_dir_L0.01_mean           0.090\n",
      "H_L0.1_mean                 0.064\n",
      "MI_dir_L0.01_variance       0.054\n",
      "H_L0.01_variance            0.043\n",
      "MI_dir_L0.1_mean            0.036\n",
      "H_L0.1_weight               0.032\n",
      "MI_dir_L0.1_weight          0.029\n",
      "HH_L0.1_magnitude           0.025\n",
      "MI_dir_L1_mean              0.025\n",
      "MI_dir_L0.1_variance        0.024\n",
      "H_L0.01_weight              0.022\n",
      "H_L1_variance               0.021\n",
      "MI_dir_L1_weight            0.020\n",
      "H_L1_mean                   0.019\n",
      "MI_dir_L0.01_weight         0.018\n",
      "MI_dir_L3_mean              0.018\n",
      "HH_L1_magnitude             0.017\n",
      "H_L0.1_variance             0.016\n",
      "MI_dir_L1_variance          0.014\n",
      "Random Forest: Accuracy=0.898\n",
      "Random Forest: f1-score=0.862\n",
      "Random Forest: Confusion Matrix [[14055     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 16314     2     2     0     0     0     0     0     0     0]\n",
      " [    0     2  8620     1     0     0     0     0     0     0     0]\n",
      " [    0     0     1  8210     1     1     0     0     0     0     0]\n",
      " [    0     0     1     1    19 26429     0     0     0     0     0]\n",
      " [    3     0     1     1     2 30922     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 33471     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 13876     0     0     1]\n",
      " [    0     0     0     0     0     0     0     0 37909     0     0]\n",
      " [    0     0     0     0     0     0     0     0     0 45602     1]\n",
      " [    0     0     0     0     0     0     0     0     0     1 23467]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     14055\n",
      "  gafgyt_combo       1.00      1.00      1.00     16318\n",
      "   gafgyt_junk       1.00      1.00      1.00      8623\n",
      "   gafgyt_scan       1.00      1.00      1.00      8213\n",
      "    gafgyt_tcp       0.86      0.00      0.00     26450\n",
      "    gafgyt_udp       0.54      1.00      0.70     30929\n",
      "     mirai_ack       1.00      1.00      1.00     33471\n",
      "    mirai_scan       1.00      1.00      1.00     13878\n",
      "     mirai_syn       1.00      1.00      1.00     37909\n",
      "     mirai_udp       1.00      1.00      1.00     45603\n",
      "mirai_udpplain       1.00      1.00      1.00     23468\n",
      "\n",
      "      accuracy                           0.90    258917\n",
      "     macro avg       0.95      0.91      0.88    258917\n",
      "  weighted avg       0.93      0.90      0.86    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.094\n",
      "MI_dir_L0.01_mean           0.077\n",
      "MI_dir_L0.1_mean            0.062\n",
      "H_L0.01_variance            0.058\n",
      "H_L0.1_mean                 0.051\n",
      "MI_dir_L0.01_variance       0.037\n",
      "H_L0.1_weight               0.034\n",
      "MI_dir_L0.1_weight          0.029\n",
      "MI_dir_L0.1_variance        0.027\n",
      "H_L0.1_variance             0.027\n",
      "H_L1_mean                   0.022\n",
      "H_L0.01_weight              0.020\n",
      "MI_dir_L1_variance          0.019\n",
      "MI_dir_L3_mean              0.018\n",
      "MI_dir_L1_weight            0.018\n",
      "H_L3_variance               0.017\n",
      "MI_dir_L1_mean              0.016\n",
      "H_L1_variance               0.016\n",
      "HH_L0.1_magnitude           0.015\n",
      "HH_L0.01_magnitude          0.014\n",
      "Random Forest: Accuracy=0.886\n",
      "Random Forest: f1-score=0.847\n",
      "Random Forest: Confusion Matrix [[ 5906     0     0     0     0     0     0     0     0     0     0]\n",
      " [    0 17831     0     0     1     0     0     0     0     0     0]\n",
      " [    0     5  8259     0     0     2     0     0     0     0     0]\n",
      " [    0     1     0  8569     0     0     0     0     0     0     0]\n",
      " [    0     0     0     0    28 29140     0     0     0     0     0]\n",
      " [    0     0     0     0     1 30674     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 32254     0     0     0     0]\n",
      " [    1     0     0     0     0     0     0 13086     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 36878     1     0]\n",
      " [    0     0     0     0     0     0     0     0     0 47134     0]\n",
      " [    0     0     0     0     0     0     3     0     0     0 25474]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      5906\n",
      "  gafgyt_combo       1.00      1.00      1.00     17832\n",
      "   gafgyt_junk       1.00      1.00      1.00      8266\n",
      "   gafgyt_scan       1.00      1.00      1.00      8570\n",
      "    gafgyt_tcp       0.93      0.00      0.00     29168\n",
      "    gafgyt_udp       0.51      1.00      0.68     30675\n",
      "     mirai_ack       1.00      1.00      1.00     32254\n",
      "    mirai_scan       1.00      1.00      1.00     13087\n",
      "     mirai_syn       1.00      1.00      1.00     36879\n",
      "     mirai_udp       1.00      1.00      1.00     47134\n",
      "mirai_udpplain       1.00      1.00      1.00     25477\n",
      "\n",
      "      accuracy                           0.89    255248\n",
      "     macro avg       0.95      0.91      0.88    255248\n",
      "  weighted avg       0.93      0.89      0.85    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with scaling unbalanced data\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier(dataframe_multi[k], k, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(319748, 115), (319748,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.064\n",
      "MI_dir_L0.01_mean           0.062\n",
      "H_L0.01_mean                0.061\n",
      "HH_jit_L0.01_mean           0.058\n",
      "HH_jit_L1_mean              0.054\n",
      "HH_jit_L3_mean              0.052\n",
      "HH_jit_L5_mean              0.040\n",
      "MI_dir_L0.1_mean            0.040\n",
      "H_L0.01_variance            0.037\n",
      "H_L0.1_weight               0.036\n",
      "H_L0.1_mean                 0.035\n",
      "MI_dir_L0.01_variance       0.030\n",
      "H_L0.1_variance             0.027\n",
      "MI_dir_L0.1_weight          0.027\n",
      "MI_dir_L0.1_variance        0.026\n",
      "MI_dir_L1_mean              0.024\n",
      "MI_dir_L0.01_weight         0.019\n",
      "MI_dir_L1_weight            0.018\n",
      "H_L1_weight                 0.017\n",
      "H_L0.01_weight              0.016\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[8725    0    0    0    0    1    0    0    0    0    0]\n",
      " [   0 8744    3    1    0    0    0    0    0    0    0]\n",
      " [   0    1 8720    2    0    0    0    0    0    0    0]\n",
      " [   0    0    0 8746    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 8745    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8675    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8685    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 8740    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8771    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8695    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8671]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8726\n",
      "  gafgyt_combo       1.00      1.00      1.00      8748\n",
      "   gafgyt_junk       1.00      1.00      1.00      8723\n",
      "   gafgyt_scan       1.00      1.00      1.00      8746\n",
      "    gafgyt_tcp       1.00      1.00      1.00      8745\n",
      "    gafgyt_udp       1.00      1.00      1.00      8675\n",
      "     mirai_ack       1.00      1.00      1.00      8685\n",
      "    mirai_scan       1.00      1.00      1.00      8740\n",
      "     mirai_syn       1.00      1.00      1.00      8771\n",
      "     mirai_udp       1.00      1.00      1.00      8695\n",
      "mirai_udpplain       1.00      1.00      1.00      8671\n",
      "\n",
      "      accuracy                           1.00     95925\n",
      "     macro avg       1.00      1.00      1.00     95925\n",
      "  weighted avg       1.00      1.00      1.00     95925\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(144243, 115), (144243,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.077\n",
      "HH_jit_L0.1_mean            0.077\n",
      "HH_jit_L1_mean              0.070\n",
      "HH_jit_L0.01_mean           0.070\n",
      "HH_jit_L3_mean              0.056\n",
      "H_L0.01_mean                0.043\n",
      "H_L0.1_mean                 0.042\n",
      "MI_dir_L0.01_mean           0.038\n",
      "MI_dir_L0.1_mean            0.036\n",
      "H_L0.1_weight               0.034\n",
      "H_L0.01_variance            0.033\n",
      "MI_dir_L0.01_variance       0.030\n",
      "HH_jit_L0.01_weight         0.025\n",
      "H_L0.01_weight              0.025\n",
      "HH_L0.01_weight             0.022\n",
      "MI_dir_L0.01_weight         0.021\n",
      "MI_dir_L0.1_weight          0.020\n",
      "MI_dir_L1_weight            0.018\n",
      "H_L0.1_variance             0.015\n",
      "H_L1_weight                 0.014\n",
      "Random Forest: Accuracy=0.989\n",
      "Random Forest: f1-score=0.988\n",
      "Random Forest: Confusion Matrix [[3932    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 3954    1    0    0    0    0    0    0    0    0]\n",
      " [   0    1 4009    0    0    0    0    0    0    0    0]\n",
      " [   1    0    0 3833    0    0    0    0    0    0    0]\n",
      " [   0    1    0    0 3436  491    0    0    0    0    0]\n",
      " [   0    1    0    0    0 3965    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 3905    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 3946    0    0    0]\n",
      " [   0    0    0    0    0    0    0    1 3905    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 3967    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 3924]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      3932\n",
      "  gafgyt_combo       1.00      1.00      1.00      3955\n",
      "   gafgyt_junk       1.00      1.00      1.00      4010\n",
      "   gafgyt_scan       1.00      1.00      1.00      3834\n",
      "    gafgyt_tcp       1.00      0.87      0.93      3928\n",
      "    gafgyt_udp       0.89      1.00      0.94      3966\n",
      "     mirai_ack       1.00      1.00      1.00      3905\n",
      "    mirai_scan       1.00      1.00      1.00      3946\n",
      "     mirai_syn       1.00      1.00      1.00      3906\n",
      "     mirai_udp       1.00      1.00      1.00      3967\n",
      "mirai_udpplain       1.00      1.00      1.00      3924\n",
      "\n",
      "      accuracy                           0.99     43273\n",
      "     macro avg       0.99      0.99      0.99     43273\n",
      "  weighted avg       0.99      0.99      0.99     43273\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(168720, 115), (168720,)\n",
      "                      importance\n",
      "feature                         \n",
      "HH_jit_L0.01_mean          0.065\n",
      "HH_jit_L0.1_mean           0.060\n",
      "HH_jit_L5_mean             0.052\n",
      "MI_dir_L0.01_weight        0.049\n",
      "HH_jit_L1_mean             0.047\n",
      "H_L0.01_weight             0.047\n",
      "MI_dir_L0.1_weight         0.045\n",
      "HH_L0.01_weight            0.043\n",
      "H_L0.1_weight              0.040\n",
      "HH_jit_L3_mean             0.040\n",
      "HH_L0.1_magnitude          0.037\n",
      "HH_L0.01_magnitude         0.035\n",
      "H_L0.01_mean               0.030\n",
      "HH_jit_L0.1_variance       0.028\n",
      "HH_jit_L0.01_weight        0.025\n",
      "MI_dir_L0.01_mean          0.021\n",
      "HH_L1_weight               0.021\n",
      "MI_dir_L1_weight           0.019\n",
      "HH_L0.1_weight             0.019\n",
      "HH_L3_weight               0.015\n",
      "Random Forest: Accuracy=0.983\n",
      "Random Forest: f1-score=0.983\n",
      "Random Forest: Confusion Matrix [[8433    0    0    0    0    0]\n",
      " [   0 8414    2    0    0    0]\n",
      " [   0    0 8409    0    0    0]\n",
      " [   0    0    1 8441    0    0]\n",
      " [   0    0    1    0 7537  849]\n",
      " [   0    0    1    0    0 8528]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8433\n",
      "gafgyt_combo       1.00      1.00      1.00      8416\n",
      " gafgyt_junk       1.00      1.00      1.00      8409\n",
      " gafgyt_scan       1.00      1.00      1.00      8442\n",
      "  gafgyt_tcp       1.00      0.90      0.95      8387\n",
      "  gafgyt_udp       0.91      1.00      0.95      8529\n",
      "\n",
      "    accuracy                           0.98     50616\n",
      "   macro avg       0.98      0.98      0.98     50616\n",
      "weighted avg       0.98      0.98      0.98     50616\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(306449, 115), (306449,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.068\n",
      "HH_jit_L0.01_mean           0.066\n",
      "HH_jit_L1_mean              0.054\n",
      "HH_jit_L3_mean              0.050\n",
      "HH_jit_L5_mean              0.049\n",
      "MI_dir_L0.01_mean           0.043\n",
      "H_L0.01_mean                0.036\n",
      "H_L0.1_mean                 0.033\n",
      "MI_dir_L0.1_mean            0.032\n",
      "H_L0.01_weight              0.031\n",
      "MI_dir_L0.1_weight          0.030\n",
      "H_L0.1_weight               0.029\n",
      "MI_dir_L0.01_weight         0.028\n",
      "H_L1_mean                   0.025\n",
      "H_L0.1_variance             0.023\n",
      "HH_L0.01_weight             0.022\n",
      "H_L0.01_variance            0.022\n",
      "H_L1_weight                 0.021\n",
      "MI_dir_L1_mean              0.018\n",
      "MI_dir_L0.01_variance       0.017\n",
      "Random Forest: Accuracy=0.989\n",
      "Random Forest: f1-score=0.989\n",
      "Random Forest: Confusion Matrix [[8339    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8407    2    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8324    0    0    0    0    0    1    0    0]\n",
      " [   2    0    1 8386    0    0    0    0    0    0    0]\n",
      " [   0    0    1    0 7456  961    0    0    0    0    0]\n",
      " [   1    0    1    0    0 8322    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8440    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 8455    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8309    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8279    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8248]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8339\n",
      "  gafgyt_combo       1.00      1.00      1.00      8409\n",
      "   gafgyt_junk       1.00      1.00      1.00      8325\n",
      "   gafgyt_scan       1.00      1.00      1.00      8389\n",
      "    gafgyt_tcp       1.00      0.89      0.94      8418\n",
      "    gafgyt_udp       0.90      1.00      0.95      8324\n",
      "     mirai_ack       1.00      1.00      1.00      8440\n",
      "    mirai_scan       1.00      1.00      1.00      8455\n",
      "     mirai_syn       1.00      1.00      1.00      8309\n",
      "     mirai_udp       1.00      1.00      1.00      8279\n",
      "mirai_udpplain       1.00      1.00      1.00      8248\n",
      "\n",
      "      accuracy                           0.99     91935\n",
      "     macro avg       0.99      0.99      0.99     91935\n",
      "  weighted avg       0.99      0.99      0.99     91935\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(322267, 115), (322267,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.060\n",
      "MI_dir_L0.01_mean           0.059\n",
      "H_L0.01_variance            0.051\n",
      "H_L0.01_mean                0.049\n",
      "MI_dir_L0.01_variance       0.048\n",
      "HH_jit_L0.1_mean            0.045\n",
      "MI_dir_L0.1_weight          0.041\n",
      "MI_dir_L0.1_mean            0.041\n",
      "H_L0.1_mean                 0.040\n",
      "H_L0.1_weight               0.039\n",
      "H_L0.1_variance             0.038\n",
      "HH_jit_L3_mean              0.038\n",
      "MI_dir_L0.1_variance        0.037\n",
      "HH_jit_L5_mean              0.036\n",
      "H_L0.01_weight              0.032\n",
      "HH_L0.01_weight             0.029\n",
      "HH_jit_L1_mean              0.027\n",
      "HH_jit_L0.01_weight         0.026\n",
      "MI_dir_L0.01_weight         0.021\n",
      "MI_dir_L1_mean              0.014\n",
      "Random Forest: Accuracy=0.981\n",
      "Random Forest: f1-score=0.981\n",
      "Random Forest: Confusion Matrix [[8795    0    0    0    0    1    0    0    0    0    0]\n",
      " [   0 8803    5    0    0    0    0    0    0    0    0]\n",
      " [   0    1 8789    0    1    0    0    0    0    0    0]\n",
      " [   0    0    0 8763    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 7038 1807    0    0    0    0    0]\n",
      " [   0    0    0    0    1 8729    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8901    0    1    0    0]\n",
      " [   0    0    1    0    0    0    0 8770    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8827    0    0]\n",
      " [   0    0    0    0    0    0    0    0    2 8673    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8773]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8796\n",
      "  gafgyt_combo       1.00      1.00      1.00      8808\n",
      "   gafgyt_junk       1.00      1.00      1.00      8791\n",
      "   gafgyt_scan       1.00      1.00      1.00      8763\n",
      "    gafgyt_tcp       1.00      0.80      0.89      8845\n",
      "    gafgyt_udp       0.83      1.00      0.91      8730\n",
      "     mirai_ack       1.00      1.00      1.00      8902\n",
      "    mirai_scan       1.00      1.00      1.00      8771\n",
      "     mirai_syn       1.00      1.00      1.00      8827\n",
      "     mirai_udp       1.00      1.00      1.00      8675\n",
      "mirai_udpplain       1.00      1.00      1.00      8773\n",
      "\n",
      "      accuracy                           0.98     96681\n",
      "     macro avg       0.98      0.98      0.98     96681\n",
      "  weighted avg       0.98      0.98      0.98     96681\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(312367, 115), (312367,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.060\n",
      "MI_dir_L0.01_mean           0.059\n",
      "H_L0.01_mean                0.053\n",
      "H_L0.1_mean                 0.046\n",
      "HH_jit_L3_mean              0.045\n",
      "HH_jit_L1_mean              0.044\n",
      "H_L0.01_variance            0.044\n",
      "MI_dir_L0.1_mean            0.042\n",
      "MI_dir_L0.01_variance       0.041\n",
      "HH_jit_L0.1_mean            0.038\n",
      "HH_jit_L5_mean              0.037\n",
      "H_L0.1_weight               0.036\n",
      "H_L0.1_variance             0.032\n",
      "H_L0.01_weight              0.032\n",
      "MI_dir_L0.01_weight         0.027\n",
      "MI_dir_L0.1_variance        0.026\n",
      "MI_dir_L0.1_weight          0.025\n",
      "HH_jit_L0.01_weight         0.022\n",
      "HH_jit_L0.1_weight          0.021\n",
      "MI_dir_L1_mean              0.017\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[8520    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8578    6    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8542    0    0    0    0    0    0    0    0]\n",
      " [   0    0    1 8545    0    0    0    0    0    0    0]\n",
      " [   0    0    1    0 8617    2    0    0    0    0    0]\n",
      " [   0    0    1    0    0 8540    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8471    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 8552    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8447    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8435    0]\n",
      " [   1    0    0    0    0    0    0    0    0    0 8452]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8520\n",
      "  gafgyt_combo       1.00      1.00      1.00      8584\n",
      "   gafgyt_junk       1.00      1.00      1.00      8542\n",
      "   gafgyt_scan       1.00      1.00      1.00      8546\n",
      "    gafgyt_tcp       1.00      1.00      1.00      8620\n",
      "    gafgyt_udp       1.00      1.00      1.00      8541\n",
      "     mirai_ack       1.00      1.00      1.00      8471\n",
      "    mirai_scan       1.00      1.00      1.00      8552\n",
      "     mirai_syn       1.00      1.00      1.00      8447\n",
      "     mirai_udp       1.00      1.00      1.00      8435\n",
      "mirai_udpplain       1.00      1.00      1.00      8453\n",
      "\n",
      "      accuracy                           1.00     93711\n",
      "     macro avg       1.00      1.00      1.00     93711\n",
      "  weighted avg       1.00      1.00      1.00     93711\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(166188, 115), (166188,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.071\n",
      "HH_jit_L0.1_mean            0.067\n",
      "HH_jit_L5_mean              0.056\n",
      "MI_dir_L0.01_weight         0.054\n",
      "HH_jit_L1_mean              0.051\n",
      "H_L0.01_mean                0.049\n",
      "H_L0.01_weight              0.045\n",
      "H_L0.1_mean                 0.045\n",
      "HH_jit_L3_mean              0.044\n",
      "MI_dir_L0.01_mean           0.035\n",
      "HH_jit_L0.01_weight         0.033\n",
      "H_L0.1_weight               0.033\n",
      "MI_dir_L0.1_weight          0.031\n",
      "MI_dir_L0.1_mean            0.031\n",
      "HH_jit_L0.1_weight          0.030\n",
      "HH_L0.01_weight             0.024\n",
      "HH_jit_L0.01_variance       0.021\n",
      "HH_jit_L0.1_variance        0.019\n",
      "MI_dir_L1_weight            0.017\n",
      "HH_L0.1_weight              0.014\n",
      "Random Forest: Accuracy=0.998\n",
      "Random Forest: f1-score=0.998\n",
      "Random Forest: Confusion Matrix [[8293    0    0    0    0    1]\n",
      " [   0 8279    3    1    1    0]\n",
      " [   0    0 8261    1    1    0]\n",
      " [   1    0    0 8416    0    0]\n",
      " [   0    0    0    0 8203   94]\n",
      " [   0    0    0    0    1 8301]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8294\n",
      "gafgyt_combo       1.00      1.00      1.00      8284\n",
      " gafgyt_junk       1.00      1.00      1.00      8263\n",
      " gafgyt_scan       1.00      1.00      1.00      8417\n",
      "  gafgyt_tcp       1.00      0.99      0.99      8297\n",
      "  gafgyt_udp       0.99      1.00      0.99      8302\n",
      "\n",
      "    accuracy                           1.00     49857\n",
      "   macro avg       1.00      1.00      1.00     49857\n",
      "weighted avg       1.00      1.00      1.00     49857\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(306075, 115), (306075,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.063\n",
      "HH_jit_L0.01_mean           0.055\n",
      "H_L0.01_mean                0.054\n",
      "MI_dir_L0.01_mean           0.051\n",
      "HH_jit_L5_mean              0.050\n",
      "HH_jit_L1_mean              0.047\n",
      "HH_jit_L3_mean              0.044\n",
      "H_L0.01_variance            0.043\n",
      "H_L0.1_mean                 0.042\n",
      "MI_dir_L0.01_variance       0.039\n",
      "MI_dir_L0.1_mean            0.034\n",
      "H_L0.1_weight               0.033\n",
      "MI_dir_L0.1_weight          0.028\n",
      "H_L1_weight                 0.026\n",
      "MI_dir_L0.01_weight         0.024\n",
      "H_L0.01_weight              0.023\n",
      "HH_L0.01_weight             0.019\n",
      "MI_dir_L1_weight            0.019\n",
      "H_L0.1_variance             0.018\n",
      "MI_dir_L0.1_variance        0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[8335    0    0    0    0    1    0    0    0    0    0]\n",
      " [   0 8401    2    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8311    0    0    0    0    0    0    0    0]\n",
      " [   1    0    0 8394    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 8461    4    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8315    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8330    0    0    0    0]\n",
      " [   0    1    0    0    0    0    0 8474    0    0    1]\n",
      " [   0    0    0    0    0    0    0    0 8231    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8314    1]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8246]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8336\n",
      "  gafgyt_combo       1.00      1.00      1.00      8403\n",
      "   gafgyt_junk       1.00      1.00      1.00      8311\n",
      "   gafgyt_scan       1.00      1.00      1.00      8395\n",
      "    gafgyt_tcp       1.00      1.00      1.00      8465\n",
      "    gafgyt_udp       1.00      1.00      1.00      8315\n",
      "     mirai_ack       1.00      1.00      1.00      8330\n",
      "    mirai_scan       1.00      1.00      1.00      8476\n",
      "     mirai_syn       1.00      1.00      1.00      8231\n",
      "     mirai_udp       1.00      1.00      1.00      8315\n",
      "mirai_udpplain       1.00      1.00      1.00      8246\n",
      "\n",
      "      accuracy                           1.00     91823\n",
      "     macro avg       1.00      1.00      1.00     91823\n",
      "  weighted avg       1.00      1.00      1.00     91823\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(214808, 115), (214808,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean              0.069\n",
      "HH_jit_L0.01_mean           0.067\n",
      "HH_jit_L3_mean              0.067\n",
      "HH_jit_L0.1_mean            0.065\n",
      "HH_jit_L5_mean              0.056\n",
      "H_L0.01_mean                0.050\n",
      "MI_dir_L0.01_mean           0.046\n",
      "MI_dir_L0.1_weight          0.034\n",
      "H_L0.1_weight               0.033\n",
      "H_L0.01_variance            0.032\n",
      "HH_jit_L0.01_weight         0.028\n",
      "MI_dir_L0.01_variance       0.027\n",
      "H_L0.1_variance             0.024\n",
      "H_L0.1_mean                 0.024\n",
      "MI_dir_L0.1_mean            0.023\n",
      "HH_L0.01_weight             0.022\n",
      "H_L0.01_weight              0.021\n",
      "MI_dir_L0.01_weight         0.020\n",
      "H_L1_weight                 0.019\n",
      "MI_dir_L1_weight            0.018\n",
      "Random Forest: Accuracy=0.971\n",
      "Random Forest: f1-score=0.970\n",
      "Random Forest: Confusion Matrix [[5904    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 5818    0    0    0    0    0    0    0    0    0]\n",
      " [   0    2 5848    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0 5758    0    1    0    0    0    0    0]\n",
      " [   0    0    0    0 3978 1852    0    0    0    0    0]\n",
      " [   1    0    0    0    1 5840    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 5953    0    0    0    0]\n",
      " [   0    2    0    0    0    0    0 5875    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 5782    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 5918    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 5910]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      5904\n",
      "  gafgyt_combo       1.00      1.00      1.00      5818\n",
      "   gafgyt_junk       1.00      1.00      1.00      5850\n",
      "   gafgyt_scan       1.00      1.00      1.00      5759\n",
      "    gafgyt_tcp       1.00      0.68      0.81      5830\n",
      "    gafgyt_udp       0.76      1.00      0.86      5842\n",
      "     mirai_ack       1.00      1.00      1.00      5953\n",
      "    mirai_scan       1.00      1.00      1.00      5877\n",
      "     mirai_syn       1.00      1.00      1.00      5782\n",
      "     mirai_udp       1.00      1.00      1.00      5918\n",
      "mirai_udpplain       1.00      1.00      1.00      5910\n",
      "\n",
      "      accuracy                           0.97     64443\n",
      "     macro avg       0.98      0.97      0.97     64443\n",
      "  weighted avg       0.98      0.97      0.97     64443\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with out scaling and balanced data\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier_balanced(dataframe_multi[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(319748, 115), (319748,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.066\n",
      "MI_dir_L0.01_mean           0.059\n",
      "HH_jit_L0.01_mean           0.055\n",
      "HH_jit_L3_mean              0.043\n",
      "HH_jit_L1_mean              0.042\n",
      "MI_dir_L0.01_variance       0.040\n",
      "H_L0.01_variance            0.038\n",
      "H_L0.1_variance             0.037\n",
      "H_L0.1_mean                 0.036\n",
      "MI_dir_L0.1_mean            0.034\n",
      "HH_jit_L0.1_mean            0.032\n",
      "HH_jit_L5_mean              0.032\n",
      "MI_dir_L0.1_weight          0.027\n",
      "H_L1_mean                   0.026\n",
      "H_L0.1_weight               0.024\n",
      "MI_dir_L0.1_variance        0.023\n",
      "H_L0.01_weight              0.023\n",
      "MI_dir_L1_mean              0.019\n",
      "HH_jit_L0.01_weight         0.019\n",
      "H_L1_weight                 0.019\n",
      "Random Forest: Accuracy=0.909\n",
      "Random Forest: f1-score=0.878\n",
      "Random Forest: Confusion Matrix [[8725    0    0    0    0    1    0    0    0    0    0]\n",
      " [   0 8744    3    1    0    0    0    0    0    0    0]\n",
      " [   0    1 8720    2    0    0    0    0    0    0    0]\n",
      " [   0    0    0 8746    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8745    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8675    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8685    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 8740    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8771    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8695    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8671]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8726\n",
      "  gafgyt_combo       1.00      1.00      1.00      8748\n",
      "   gafgyt_junk       1.00      1.00      1.00      8723\n",
      "   gafgyt_scan       1.00      1.00      1.00      8746\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8745\n",
      "    gafgyt_udp       0.50      1.00      0.66      8675\n",
      "     mirai_ack       1.00      1.00      1.00      8685\n",
      "    mirai_scan       1.00      1.00      1.00      8740\n",
      "     mirai_syn       1.00      1.00      1.00      8771\n",
      "     mirai_udp       1.00      1.00      1.00      8695\n",
      "mirai_udpplain       1.00      1.00      1.00      8671\n",
      "\n",
      "      accuracy                           0.91     95925\n",
      "     macro avg       0.86      0.91      0.88     95925\n",
      "  weighted avg       0.86      0.91      0.88     95925\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(144243, 115), (144243,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.085\n",
      "HH_jit_L1_mean              0.079\n",
      "HH_jit_L3_mean              0.064\n",
      "HH_jit_L0.01_mean           0.062\n",
      "HH_jit_L0.1_mean            0.056\n",
      "MI_dir_L0.01_mean           0.055\n",
      "H_L0.01_mean                0.047\n",
      "H_L0.1_mean                 0.032\n",
      "MI_dir_L0.1_weight          0.029\n",
      "MI_dir_L0.1_mean            0.028\n",
      "H_L0.1_weight               0.026\n",
      "H_L0.01_weight              0.025\n",
      "MI_dir_L0.01_variance       0.024\n",
      "H_L0.01_variance            0.023\n",
      "MI_dir_L0.01_weight         0.023\n",
      "HH_L0.01_weight             0.023\n",
      "MI_dir_L1_weight            0.018\n",
      "HH_jit_L0.01_weight         0.015\n",
      "H_L1_weight                 0.014\n",
      "H_L1_mean                   0.013\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[3932    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 3954    1    0    0    0    0    0    0    0    0]\n",
      " [   0    1 4009    0    0    0    0    0    0    0    0]\n",
      " [   1    0    0 3833    0    0    0    0    0    0    0]\n",
      " [   0    1    0    0 3926    1    0    0    0    0    0]\n",
      " [   0    1    0    0    0 3965    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 3905    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 3946    0    0    0]\n",
      " [   0    0    0    0    0    0    0    1 3905    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 3967    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 3924]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      3932\n",
      "  gafgyt_combo       1.00      1.00      1.00      3955\n",
      "   gafgyt_junk       1.00      1.00      1.00      4010\n",
      "   gafgyt_scan       1.00      1.00      1.00      3834\n",
      "    gafgyt_tcp       1.00      1.00      1.00      3928\n",
      "    gafgyt_udp       1.00      1.00      1.00      3966\n",
      "     mirai_ack       1.00      1.00      1.00      3905\n",
      "    mirai_scan       1.00      1.00      1.00      3946\n",
      "     mirai_syn       1.00      1.00      1.00      3906\n",
      "     mirai_udp       1.00      1.00      1.00      3967\n",
      "mirai_udpplain       1.00      1.00      1.00      3924\n",
      "\n",
      "      accuracy                           1.00     43273\n",
      "     macro avg       1.00      1.00      1.00     43273\n",
      "  weighted avg       1.00      1.00      1.00     43273\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(168720, 115), (168720,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.076\n",
      "MI_dir_L0.1_weight          0.062\n",
      "H_L0.01_weight              0.061\n",
      "H_L0.01_mean                0.054\n",
      "MI_dir_L0.01_weight         0.053\n",
      "MI_dir_L0.1_mean            0.031\n",
      "MI_dir_L0.01_mean           0.031\n",
      "HH_jit_L0.01_weight         0.029\n",
      "HH_L0.01_weight             0.029\n",
      "HH_jit_L0.01_mean           0.024\n",
      "HpHp_L5_magnitude           0.023\n",
      "HH_L0.1_magnitude           0.022\n",
      "HH_L3_magnitude             0.022\n",
      "HpHp_L0.01_magnitude        0.021\n",
      "HH_jit_L0.01_variance       0.021\n",
      "HH_L5_weight                0.020\n",
      "HH_L0.01_magnitude          0.019\n",
      "HH_L1_weight                0.018\n",
      "HH_L5_magnitude             0.018\n",
      "HH_jit_L1_weight            0.018\n",
      "Random Forest: Accuracy=0.831\n",
      "Random Forest: f1-score=0.776\n",
      "Random Forest: Confusion Matrix [[8433    0    0    0    0    0]\n",
      " [   0 8414    2    0    0    0]\n",
      " [   0    0 8409    0    0    0]\n",
      " [   0    0    1 8441    0    0]\n",
      " [   0    0    1    0 8386    0]\n",
      " [   0    0    1    0 8528    0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8433\n",
      "gafgyt_combo       1.00      1.00      1.00      8416\n",
      " gafgyt_junk       1.00      1.00      1.00      8409\n",
      " gafgyt_scan       1.00      1.00      1.00      8442\n",
      "  gafgyt_tcp       0.50      1.00      0.66      8387\n",
      "  gafgyt_udp       0.00      0.00      0.00      8529\n",
      "\n",
      "    accuracy                           0.83     50616\n",
      "   macro avg       0.75      0.83      0.78     50616\n",
      "weighted avg       0.75      0.83      0.78     50616\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(306449, 115), (306449,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.096\n",
      "HH_jit_L1_mean              0.075\n",
      "HH_jit_L5_mean              0.043\n",
      "H_L0.1_weight               0.040\n",
      "H_L0.01_mean                0.040\n",
      "HH_jit_L0.1_mean            0.039\n",
      "H_L0.01_weight              0.036\n",
      "MI_dir_L0.1_weight          0.035\n",
      "H_L0.1_mean                 0.032\n",
      "HH_jit_L3_mean              0.031\n",
      "MI_dir_L0.01_variance       0.029\n",
      "MI_dir_L0.01_mean           0.029\n",
      "MI_dir_L0.01_weight         0.028\n",
      "MI_dir_L0.1_mean            0.025\n",
      "H_L0.1_variance             0.025\n",
      "H_L0.01_variance            0.025\n",
      "HH_jit_L0.01_weight         0.020\n",
      "MI_dir_L0.1_variance        0.019\n",
      "H_L1_mean                   0.016\n",
      "HH_L0.01_weight             0.015\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[8339    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8408    1    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8325    0    0    0    0    0    0    0    0]\n",
      " [   1    0    1 8387    0    0    0    0    0    0    0]\n",
      " [   0    0    1    0 8417    0    0    0    0    0    0]\n",
      " [   1    0    1    0    0 8322    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8440    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 8455    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8309    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8279    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8248]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8339\n",
      "  gafgyt_combo       1.00      1.00      1.00      8409\n",
      "   gafgyt_junk       1.00      1.00      1.00      8325\n",
      "   gafgyt_scan       1.00      1.00      1.00      8389\n",
      "    gafgyt_tcp       1.00      1.00      1.00      8418\n",
      "    gafgyt_udp       1.00      1.00      1.00      8324\n",
      "     mirai_ack       1.00      1.00      1.00      8440\n",
      "    mirai_scan       1.00      1.00      1.00      8455\n",
      "     mirai_syn       1.00      1.00      1.00      8309\n",
      "     mirai_udp       1.00      1.00      1.00      8279\n",
      "mirai_udpplain       1.00      1.00      1.00      8248\n",
      "\n",
      "      accuracy                           1.00     91935\n",
      "     macro avg       1.00      1.00      1.00     91935\n",
      "  weighted avg       1.00      1.00      1.00     91935\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(322267, 115), (322267,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.064\n",
      "H_L0.01_mean                0.052\n",
      "MI_dir_L0.1_mean            0.051\n",
      "MI_dir_L0.01_variance       0.047\n",
      "H_L0.1_mean                 0.047\n",
      "H_L0.1_weight               0.046\n",
      "MI_dir_L0.1_variance        0.042\n",
      "HH_jit_L0.1_mean            0.041\n",
      "HH_jit_L0.01_mean           0.039\n",
      "H_L0.1_variance             0.038\n",
      "H_L0.01_variance            0.035\n",
      "MI_dir_L0.1_weight          0.034\n",
      "HH_jit_L5_mean              0.033\n",
      "HH_jit_L1_mean              0.032\n",
      "HH_jit_L0.01_weight         0.031\n",
      "H_L0.01_weight              0.030\n",
      "HH_jit_L3_mean              0.026\n",
      "HH_L0.01_weight             0.022\n",
      "MI_dir_L0.01_weight         0.022\n",
      "HH_jit_L1_weight            0.014\n",
      "Random Forest: Accuracy=0.908\n",
      "Random Forest: f1-score=0.878\n",
      "Random Forest: Confusion Matrix [[8796    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8803    5    0    0    0    0    0    0    0    0]\n",
      " [   0    2 8789    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0 8763    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8845    0    0    0    0    0]\n",
      " [   0    0    0    0    1 8729    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8901    0    1    0    0]\n",
      " [   0    0    1    0    0    0    0 8770    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8827    0    0]\n",
      " [   0    0    0    0    0    0    0    0    2 8673    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8773]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8796\n",
      "  gafgyt_combo       1.00      1.00      1.00      8808\n",
      "   gafgyt_junk       1.00      1.00      1.00      8791\n",
      "   gafgyt_scan       1.00      1.00      1.00      8763\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8845\n",
      "    gafgyt_udp       0.50      1.00      0.66      8730\n",
      "     mirai_ack       1.00      1.00      1.00      8902\n",
      "    mirai_scan       1.00      1.00      1.00      8771\n",
      "     mirai_syn       1.00      1.00      1.00      8827\n",
      "     mirai_udp       1.00      1.00      1.00      8675\n",
      "mirai_udpplain       1.00      1.00      1.00      8773\n",
      "\n",
      "      accuracy                           0.91     96681\n",
      "     macro avg       0.86      0.91      0.88     96681\n",
      "  weighted avg       0.86      0.91      0.88     96681\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(312367, 115), (312367,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.067\n",
      "H_L0.01_mean                0.059\n",
      "HH_jit_L1_mean              0.054\n",
      "MI_dir_L0.01_variance       0.050\n",
      "MI_dir_L0.1_mean            0.048\n",
      "H_L0.01_variance            0.048\n",
      "MI_dir_L0.01_mean           0.043\n",
      "HH_jit_L0.01_mean           0.043\n",
      "H_L0.1_mean                 0.041\n",
      "MI_dir_L0.1_weight          0.033\n",
      "H_L0.1_weight               0.033\n",
      "H_L0.1_variance             0.032\n",
      "MI_dir_L0.01_weight         0.030\n",
      "HH_L0.01_weight             0.029\n",
      "MI_dir_L0.1_variance        0.029\n",
      "H_L0.01_weight              0.025\n",
      "HH_jit_L0.01_weight         0.024\n",
      "H_L1_mean                   0.019\n",
      "MI_dir_L1_mean              0.019\n",
      "HH_jit_L3_mean              0.018\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[8520    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8578    6    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8542    0    0    0    0    0    0    0    0]\n",
      " [   0    0    1 8545    0    0    0    0    0    0    0]\n",
      " [   0    0    1    0 8619    0    0    0    0    0    0]\n",
      " [   0    0    1    0    0 8540    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8471    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 8552    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8447    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8435    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8453]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8520\n",
      "  gafgyt_combo       1.00      1.00      1.00      8584\n",
      "   gafgyt_junk       1.00      1.00      1.00      8542\n",
      "   gafgyt_scan       1.00      1.00      1.00      8546\n",
      "    gafgyt_tcp       1.00      1.00      1.00      8620\n",
      "    gafgyt_udp       1.00      1.00      1.00      8541\n",
      "     mirai_ack       1.00      1.00      1.00      8471\n",
      "    mirai_scan       1.00      1.00      1.00      8552\n",
      "     mirai_syn       1.00      1.00      1.00      8447\n",
      "     mirai_udp       1.00      1.00      1.00      8435\n",
      "mirai_udpplain       1.00      1.00      1.00      8453\n",
      "\n",
      "      accuracy                           1.00     93711\n",
      "     macro avg       1.00      1.00      1.00     93711\n",
      "  weighted avg       1.00      1.00      1.00     93711\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(166188, 115), (166188,)\n",
      "                      importance\n",
      "feature                         \n",
      "MI_dir_L0.01_mean          0.081\n",
      "MI_dir_L0.01_weight        0.068\n",
      "MI_dir_L0.1_mean           0.051\n",
      "MI_dir_L0.1_weight         0.049\n",
      "H_L0.1_mean                0.045\n",
      "H_L0.1_weight              0.045\n",
      "H_L0.01_mean               0.043\n",
      "HH_jit_L0.01_mean          0.043\n",
      "HH_jit_L0.01_weight        0.040\n",
      "HH_L0.01_weight            0.038\n",
      "H_L0.01_weight             0.034\n",
      "H_L1_weight                0.022\n",
      "MI_dir_L1_mean             0.022\n",
      "HH_L0.1_weight             0.021\n",
      "HH_L0.1_magnitude          0.018\n",
      "HH_L1_weight               0.018\n",
      "HpHp_L0.1_magnitude        0.018\n",
      "HH_jit_L0.1_mean           0.018\n",
      "HpHp_L1_magnitude          0.018\n",
      "HpHp_L0.01_magnitude       0.016\n",
      "Random Forest: Accuracy=0.833\n",
      "Random Forest: f1-score=0.778\n",
      "Random Forest: Confusion Matrix [[8293    0    0    0    1    0]\n",
      " [   0 8279    3    1    1    0]\n",
      " [   0    0 8261    1    1    0]\n",
      " [   0    0    0 8417    0    0]\n",
      " [   0    0    0    0    0 8297]\n",
      " [   0    0    0    1    0 8301]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8294\n",
      "gafgyt_combo       1.00      1.00      1.00      8284\n",
      " gafgyt_junk       1.00      1.00      1.00      8263\n",
      " gafgyt_scan       1.00      1.00      1.00      8417\n",
      "  gafgyt_tcp       0.00      0.00      0.00      8297\n",
      "  gafgyt_udp       0.50      1.00      0.67      8302\n",
      "\n",
      "    accuracy                           0.83     49857\n",
      "   macro avg       0.75      0.83      0.78     49857\n",
      "weighted avg       0.75      0.83      0.78     49857\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(306075, 115), (306075,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.066\n",
      "MI_dir_L0.01_mean           0.065\n",
      "HH_jit_L0.01_mean           0.058\n",
      "HH_jit_L0.1_mean            0.055\n",
      "MI_dir_L0.01_variance       0.053\n",
      "H_L0.1_mean                 0.041\n",
      "H_L0.1_weight               0.040\n",
      "HH_jit_L1_mean              0.039\n",
      "MI_dir_L0.1_mean            0.037\n",
      "H_L0.01_variance            0.036\n",
      "HH_jit_L5_mean              0.035\n",
      "HH_jit_L3_mean              0.034\n",
      "MI_dir_L0.01_weight         0.034\n",
      "MI_dir_L0.1_weight          0.027\n",
      "H_L0.01_weight              0.025\n",
      "HH_jit_L0.01_weight         0.024\n",
      "MI_dir_L0.1_variance        0.019\n",
      "HH_L0.01_weight             0.019\n",
      "H_L0.1_variance             0.019\n",
      "HH_L1_weight                0.017\n",
      "Random Forest: Accuracy=0.908\n",
      "Random Forest: f1-score=0.877\n",
      "Random Forest: Confusion Matrix [[8336    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8400    3    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8311    0    0    0    0    0    0    0    0]\n",
      " [   1    0    0 8394    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8465    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8315    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8330    0    0    0    0]\n",
      " [   0    1    0    0    0    0    0 8474    0    0    1]\n",
      " [   0    0    0    0    0    0    0    0 8231    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8314    1]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8246]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8336\n",
      "  gafgyt_combo       1.00      1.00      1.00      8403\n",
      "   gafgyt_junk       1.00      1.00      1.00      8311\n",
      "   gafgyt_scan       1.00      1.00      1.00      8395\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8465\n",
      "    gafgyt_udp       0.50      1.00      0.66      8315\n",
      "     mirai_ack       1.00      1.00      1.00      8330\n",
      "    mirai_scan       1.00      1.00      1.00      8476\n",
      "     mirai_syn       1.00      1.00      1.00      8231\n",
      "     mirai_udp       1.00      1.00      1.00      8315\n",
      "mirai_udpplain       1.00      1.00      1.00      8246\n",
      "\n",
      "      accuracy                           0.91     91823\n",
      "     macro avg       0.86      0.91      0.88     91823\n",
      "  weighted avg       0.86      0.91      0.88     91823\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(214808, 115), (214808,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean              0.087\n",
      "HH_jit_L3_mean              0.075\n",
      "HH_jit_L0.1_mean            0.068\n",
      "HH_jit_L0.01_mean           0.061\n",
      "H_L0.01_mean                0.054\n",
      "HH_jit_L5_mean              0.045\n",
      "MI_dir_L0.01_mean           0.044\n",
      "MI_dir_L0.1_mean            0.041\n",
      "H_L0.1_weight               0.036\n",
      "H_L0.01_weight              0.034\n",
      "MI_dir_L0.01_variance       0.032\n",
      "MI_dir_L0.1_weight          0.030\n",
      "H_L0.1_mean                 0.028\n",
      "H_L0.01_variance            0.026\n",
      "MI_dir_L0.01_weight         0.022\n",
      "HH_L0.01_weight             0.022\n",
      "HH_jit_L0.01_weight         0.018\n",
      "MI_dir_L1_weight            0.015\n",
      "H_L0.1_variance             0.013\n",
      "HH_L1_weight                0.013\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[5904    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 5818    0    0    0    0    0    0    0    0    0]\n",
      " [   0    1 5848    0    0    1    0    0    0    0    0]\n",
      " [   0    0    0 5758    0    1    0    0    0    0    0]\n",
      " [   0    0    0    0 5829    1    0    0    0    0    0]\n",
      " [   0    0    0    1    0 5841    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 5953    0    0    0    0]\n",
      " [   0    2    0    0    0    0    0 5875    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 5782    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 5918    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 5910]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      5904\n",
      "  gafgyt_combo       1.00      1.00      1.00      5818\n",
      "   gafgyt_junk       1.00      1.00      1.00      5850\n",
      "   gafgyt_scan       1.00      1.00      1.00      5759\n",
      "    gafgyt_tcp       1.00      1.00      1.00      5830\n",
      "    gafgyt_udp       1.00      1.00      1.00      5842\n",
      "     mirai_ack       1.00      1.00      1.00      5953\n",
      "    mirai_scan       1.00      1.00      1.00      5877\n",
      "     mirai_syn       1.00      1.00      1.00      5782\n",
      "     mirai_udp       1.00      1.00      1.00      5918\n",
      "mirai_udpplain       1.00      1.00      1.00      5910\n",
      "\n",
      "      accuracy                           1.00     64443\n",
      "     macro avg       1.00      1.00      1.00     64443\n",
      "  weighted avg       1.00      1.00      1.00     64443\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with out scaling and balanced data\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    rf_classifier_balanced(dataframe_multi[k], k, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dt_classifier(data, device_name, scaling=False,):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        model_name = f'{device_name}_without_scaling_unbalanced_model_decision_tree_gini.pkl'\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        model_name = f'{device_name}_with_scaling_unbalanced_model_decision_tree_gini.pkl'\n",
    "    \n",
    "    clf = DecisionTreeClassifier(criterion='gini', random_state= 1234, max_depth  = 5 )\n",
    "    model_res = clf.fit(X_train, y_train)\n",
    "    y_pred = model_res.predict(X_test)\n",
    "    y_pred_prob = model_res.predict_proba(X_test)\n",
    "    lr_probs = y_pred_prob[:,1]\n",
    "    ac = accuracy_score(y_test, y_pred)\n",
    "    importances = pd.DataFrame({'feature':X.columns,'importance':np.round(clf.feature_importances_,3)})\n",
    "    importances = importances.sort_values('importance',ascending=False).set_index('feature')\n",
    "\n",
    "    f1 = f1_score(y_test, y_pred, average='weighted')\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "    print(importances.head(20))\n",
    "    \n",
    "    print('Random Forest: Accuracy=%.3f' % (ac))\n",
    "\n",
    "    print('Random Forest: f1-score=%.3f' % (f1))\n",
    "    print('Random Forest: Confusion Matrix', cm)\n",
    "    print('Random Forest : Classification Report',classification_report(y_test, y_pred))\n",
    "    \n",
    "    best_model = clf\n",
    "    best_model.version = 1.0\n",
    "    best_model.pandas_version = pd.__version__\n",
    "    best_model.numpy_version = np.__version__\n",
    "    best_model.sklearn_version = sklearn_version\n",
    "    #best_model.X_columns = [col for col in X_train.columns]\n",
    "    best_model.build_datetime = datetime.now()\n",
    "    \n",
    "    modelpath = 'models'\n",
    "    if not os.path.exists(modelpath):\n",
    "        os.mkdir(modelpath)\n",
    "    iotmodel_path = os.path.join(modelpath, model_name)\n",
    "    if not os.path.exists(iotmodel_path):\n",
    "        with open(iotmodel_path, 'wb') as f:\n",
    "            pickle.dump(best_model, f)\n",
    "    \n",
    "    print('---------------------- Done ---------------------')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_weight         0.558\n",
      "HH_L0.1_weight              0.238\n",
      "H_L0.01_mean                0.161\n",
      "HH_jit_L3_weight            0.011\n",
      "HH_jit_L3_mean              0.010\n",
      "MI_dir_L0.01_mean           0.007\n",
      "H_L0.1_mean                 0.006\n",
      "HH_jit_L0.01_variance       0.003\n",
      "HH_jit_L1_mean              0.003\n",
      "HH_jit_L3_variance          0.002\n",
      "HH_jit_L1_variance          0.001\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 14993      8      2]\n",
      " [    12  94772     75]\n",
      " [     1     27 195600]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15003\n",
      "      gafgyt       1.00      1.00      1.00     94859\n",
      "       mirai       1.00      1.00      1.00    195628\n",
      "\n",
      "    accuracy                           1.00    305490\n",
      "   macro avg       1.00      1.00      1.00    305490\n",
      "weighted avg       1.00      1.00      1.00    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                    importance\n",
      "feature                       \n",
      "MI_dir_L3_variance       0.619\n",
      "HH_jit_L1_mean           0.170\n",
      "MI_dir_L0.1_weight       0.061\n",
      "HH_L0.1_weight           0.057\n",
      "H_L0.01_mean             0.032\n",
      "H_L0.01_weight           0.020\n",
      "HH_L5_magnitude          0.019\n",
      "H_L3_weight              0.007\n",
      "HH_jit_L1_variance       0.007\n",
      "HH_jit_L1_weight         0.005\n",
      "HH_jit_L0.1_mean         0.002\n",
      "MI_dir_L5_weight         0.000\n",
      "HpHp_L5_weight           0.000\n",
      "HpHp_L5_std              0.000\n",
      "HpHp_L5_mean             0.000\n",
      "HH_jit_L0.01_mean        0.000\n",
      "HpHp_L5_magnitude        0.000\n",
      "HpHp_L5_radius           0.000\n",
      "HpHp_L5_covariance       0.000\n",
      "HpHp_L5_pcc              0.000\n",
      "Random Forest: Accuracy=0.997\n",
      "Random Forest: f1-score=0.997\n",
      "Random Forest: Confusion Matrix [[  3998     18      0]\n",
      " [    12  92105    689]\n",
      " [     8     58 153875]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      4016\n",
      "      gafgyt       1.00      0.99      1.00     92806\n",
      "       mirai       1.00      1.00      1.00    153941\n",
      "\n",
      "    accuracy                           1.00    250763\n",
      "   macro avg       1.00      1.00      1.00    250763\n",
      "weighted avg       1.00      1.00      1.00    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.957\n",
      "HH_jit_L0.01_mean           0.019\n",
      "HH_jit_L5_variance          0.009\n",
      "HH_L0.01_pcc                0.006\n",
      "MI_dir_L0.1_weight          0.005\n",
      "HpHp_L0.1_weight            0.001\n",
      "MI_dir_L3_mean              0.001\n",
      "HH_jit_L3_variance          0.001\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.999\n",
      "Random Forest: f1-score=0.999\n",
      "Random Forest: Confusion Matrix [[11845    58]\n",
      " [    7 94740]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "      gafgyt       1.00      1.00      1.00     94747\n",
      "\n",
      "    accuracy                           1.00    106650\n",
      "   macro avg       1.00      1.00      1.00    106650\n",
      "weighted avg       1.00      1.00      1.00    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_weight         0.474\n",
      "HpHp_L1_magnitude           0.309\n",
      "HH_L0.1_weight              0.190\n",
      "HH_jit_L1_variance          0.010\n",
      "HH_jit_L1_mean              0.008\n",
      "HH_L0.01_weight             0.004\n",
      "H_L0.1_weight               0.003\n",
      "MI_dir_L0.01_variance       0.001\n",
      "HH_L0.01_magnitude          0.001\n",
      "HH_jit_L5_variance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_L0.01_radius             0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 52472     25      1]\n",
      " [    54  93819      7]\n",
      " [     0     19 183207]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52498\n",
      "      gafgyt       1.00      1.00      1.00     93880\n",
      "       mirai       1.00      1.00      1.00    183226\n",
      "\n",
      "    accuracy                           1.00    329604\n",
      "   macro avg       1.00      1.00      1.00    329604\n",
      "weighted avg       1.00      1.00      1.00    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.460\n",
      "HH_L0.1_weight              0.308\n",
      "HH_jit_L0.1_mean            0.205\n",
      "HH_L1_weight                0.015\n",
      "MI_dir_L0.1_variance        0.008\n",
      "H_L0.01_mean                0.002\n",
      "HpHp_L5_weight              0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 18688     70      0]\n",
      " [     6  98590     10]\n",
      " [     0      0 131114]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18758\n",
      "      gafgyt       1.00      1.00      1.00     98606\n",
      "       mirai       1.00      1.00      1.00    131114\n",
      "\n",
      "    accuracy                           1.00    248478\n",
      "   macro avg       1.00      1.00      1.00    248478\n",
      "weighted avg       1.00      1.00      1.00    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.431\n",
      "HH_L0.1_weight              0.287\n",
      "HH_jit_L0.01_mean           0.273\n",
      "HH_jit_L1_weight            0.007\n",
      "HH_L5_magnitude             0.001\n",
      "HH_jit_L3_mean              0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 29679     19      0]\n",
      " [    33  92208      6]\n",
      " [     0      4 129119]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29698\n",
      "      gafgyt       1.00      1.00      1.00     92247\n",
      "       mirai       1.00      1.00      1.00    129123\n",
      "\n",
      "    accuracy                           1.00    251068\n",
      "   macro avg       1.00      1.00      1.00    251068\n",
      "weighted avg       1.00      1.00      1.00    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.997\n",
      "H_L0.01_mean                0.001\n",
      "HpHp_L0.01_magnitude        0.001\n",
      "HpHp_L0.01_weight           0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15674    10]\n",
      " [   15 96868]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "      gafgyt       1.00      1.00      1.00     96883\n",
      "\n",
      "    accuracy                           1.00    112567\n",
      "   macro avg       1.00      1.00      1.00    112567\n",
      "weighted avg       1.00      1.00      1.00    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L3_variance          0.524\n",
      "HH_jit_L1_mean              0.152\n",
      "H_L0.01_mean                0.139\n",
      "MI_dir_L0.1_weight          0.072\n",
      "HH_jit_L0.1_weight          0.048\n",
      "H_L0.01_weight              0.027\n",
      "HH_L1_magnitude             0.025\n",
      "MI_dir_L0.01_weight         0.005\n",
      "HH_jit_L0.1_mean            0.004\n",
      "HH_jit_L1_variance          0.002\n",
      "MI_dir_L0.1_mean            0.002\n",
      "MI_dir_L0.01_variance       0.001\n",
      "HpHp_L5_std                 0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "Random Forest: Accuracy=0.996\n",
      "Random Forest: f1-score=0.996\n",
      "Random Forest: Confusion Matrix [[ 14049      5      1]\n",
      " [    41  89505    987]\n",
      " [     6    107 154216]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14055\n",
      "      gafgyt       1.00      0.99      0.99     90533\n",
      "       mirai       0.99      1.00      1.00    154329\n",
      "\n",
      "    accuracy                           1.00    258917\n",
      "   macro avg       1.00      1.00      1.00    258917\n",
      "weighted avg       1.00      1.00      1.00    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L3_variance          0.594\n",
      "HH_jit_L1_mean              0.162\n",
      "H_L0.01_mean                0.113\n",
      "HH_jit_L0.1_weight          0.054\n",
      "MI_dir_L0.1_weight          0.053\n",
      "HH_jit_L1_variance          0.011\n",
      "HH_jit_L1_weight            0.006\n",
      "HH_jit_L0.1_mean            0.004\n",
      "H_L0.1_mean                 0.002\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "Random Forest: Accuracy=0.997\n",
      "Random Forest: f1-score=0.997\n",
      "Random Forest: Confusion Matrix [[  5892     14      0]\n",
      " [     0  93785    726]\n",
      " [     4    119 154708]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5906\n",
      "      gafgyt       1.00      0.99      1.00     94511\n",
      "       mirai       1.00      1.00      1.00    154831\n",
      "\n",
      "    accuracy                           1.00    255248\n",
      "   macro avg       1.00      1.00      1.00    255248\n",
      "weighted avg       1.00      1.00      1.00    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with out scaling and unbalanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier(dataframe[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_weight         0.559\n",
      "HH_L0.1_weight              0.242\n",
      "H_L0.01_mean                0.161\n",
      "HH_jit_L3_weight            0.011\n",
      "HH_jit_L3_mean              0.010\n",
      "MI_dir_L0.01_mean           0.007\n",
      "MI_dir_L0.1_variance        0.007\n",
      "H_L0.1_weight               0.002\n",
      "HH_jit_L1_variance          0.001\n",
      "HH_L0.01_pcc                0.001\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "Random Forest: Accuracy=0.999\n",
      "Random Forest: f1-score=0.999\n",
      "Random Forest: Confusion Matrix [[ 14992     11      0]\n",
      " [    12  94613    234]\n",
      " [     1      3 195624]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15003\n",
      "      gafgyt       1.00      1.00      1.00     94859\n",
      "       mirai       1.00      1.00      1.00    195628\n",
      "\n",
      "    accuracy                           1.00    305490\n",
      "   macro avg       1.00      1.00      1.00    305490\n",
      "weighted avg       1.00      1.00      1.00    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L3_variance          0.620\n",
      "HH_jit_L1_mean              0.170\n",
      "MI_dir_L0.1_weight          0.061\n",
      "HH_jit_L0.1_weight          0.057\n",
      "H_L0.01_mean                0.032\n",
      "H_L0.01_weight              0.020\n",
      "HH_L1_magnitude             0.019\n",
      "MI_dir_L3_weight            0.007\n",
      "HH_jit_L0.1_mean            0.006\n",
      "HH_jit_L1_weight            0.005\n",
      "HH_jit_L0.01_mean           0.002\n",
      "HH_jit_L0.1_variance        0.001\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_std                 0.000\n",
      "Random Forest: Accuracy=0.997\n",
      "Random Forest: f1-score=0.997\n",
      "Random Forest: Confusion Matrix [[  4007      9      0]\n",
      " [    12  92106    688]\n",
      " [     8    107 153826]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      4016\n",
      "      gafgyt       1.00      0.99      1.00     92806\n",
      "       mirai       1.00      1.00      1.00    153941\n",
      "\n",
      "    accuracy                           1.00    250763\n",
      "   macro avg       1.00      1.00      1.00    250763\n",
      "weighted avg       1.00      1.00      1.00    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.957\n",
      "HH_jit_L0.01_mean           0.019\n",
      "HH_jit_L5_variance          0.009\n",
      "HH_L0.01_pcc                0.006\n",
      "MI_dir_L0.1_weight          0.005\n",
      "MI_dir_L1_mean              0.001\n",
      "HpHp_L0.1_weight            0.001\n",
      "HH_jit_L1_variance          0.001\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_radius              0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.999\n",
      "Random Forest: f1-score=0.999\n",
      "Random Forest: Confusion Matrix [[11844    59]\n",
      " [    7 94740]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "      gafgyt       1.00      1.00      1.00     94747\n",
      "\n",
      "    accuracy                           1.00    106650\n",
      "   macro avg       1.00      1.00      1.00    106650\n",
      "weighted avg       1.00      1.00      1.00    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_weight         0.474\n",
      "HpHp_L1_magnitude           0.309\n",
      "HH_jit_L0.1_weight          0.190\n",
      "HH_L3_weight                0.009\n",
      "HH_jit_L1_mean              0.008\n",
      "H_L0.1_weight               0.003\n",
      "HH_jit_L0.1_variance        0.003\n",
      "HH_L0.01_weight             0.002\n",
      "HpHp_L3_weight              0.001\n",
      "HH_L0.1_magnitude           0.001\n",
      "HH_jit_L5_variance          0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_L0.01_radius             0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_L0.01_covariance         0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 52472     25      1]\n",
      " [    53  93756     71]\n",
      " [     0     14 183212]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52498\n",
      "      gafgyt       1.00      1.00      1.00     93880\n",
      "       mirai       1.00      1.00      1.00    183226\n",
      "\n",
      "    accuracy                           1.00    329604\n",
      "   macro avg       1.00      1.00      1.00    329604\n",
      "weighted avg       1.00      1.00      1.00    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.456\n",
      "HH_L0.1_weight              0.308\n",
      "HH_jit_L0.1_mean            0.205\n",
      "HH_jit_L1_weight            0.015\n",
      "H_L0.1_variance             0.008\n",
      "MI_dir_L0.01_weight         0.003\n",
      "MI_dir_L0.01_mean           0.002\n",
      "MI_dir_L0.1_weight          0.001\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 18688     70      0]\n",
      " [     5  98591     10]\n",
      " [     0      0 131114]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18758\n",
      "      gafgyt       1.00      1.00      1.00     98606\n",
      "       mirai       1.00      1.00      1.00    131114\n",
      "\n",
      "    accuracy                           1.00    248478\n",
      "   macro avg       1.00      1.00      1.00    248478\n",
      "weighted avg       1.00      1.00      1.00    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.430\n",
      "HH_L0.1_weight              0.287\n",
      "HH_jit_L0.01_mean           0.273\n",
      "HH_L1_weight                0.007\n",
      "HH_jit_L3_mean              0.001\n",
      "HH_L5_magnitude             0.001\n",
      "MI_dir_L0.1_weight          0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 29679     19      0]\n",
      " [    33  92208      6]\n",
      " [     0      4 129119]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29698\n",
      "      gafgyt       1.00      1.00      1.00     92247\n",
      "       mirai       1.00      1.00      1.00    129123\n",
      "\n",
      "    accuracy                           1.00    251068\n",
      "   macro avg       1.00      1.00      1.00    251068\n",
      "weighted avg       1.00      1.00      1.00    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.998\n",
      "HpHp_L0.01_magnitude        0.001\n",
      "HpHp_L0.01_weight           0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15674    10]\n",
      " [   15 96868]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "      gafgyt       1.00      1.00      1.00     96883\n",
      "\n",
      "    accuracy                           1.00    112567\n",
      "   macro avg       1.00      1.00      1.00    112567\n",
      "weighted avg       1.00      1.00      1.00    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L3_variance          0.524\n",
      "HH_jit_L1_mean              0.152\n",
      "H_L0.01_mean                0.141\n",
      "MI_dir_L0.1_weight          0.072\n",
      "HH_L0.1_weight              0.048\n",
      "MI_dir_L0.01_weight         0.027\n",
      "HH_L1_magnitude             0.025\n",
      "H_L0.01_weight              0.005\n",
      "HH_jit_L0.01_mean           0.003\n",
      "HH_jit_L0.1_variance        0.002\n",
      "MI_dir_L0.01_variance       0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "Random Forest: Accuracy=0.996\n",
      "Random Forest: f1-score=0.996\n",
      "Random Forest: Confusion Matrix [[ 14049      5      1]\n",
      " [    41  89504    988]\n",
      " [     6    118 154205]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14055\n",
      "      gafgyt       1.00      0.99      0.99     90533\n",
      "       mirai       0.99      1.00      1.00    154329\n",
      "\n",
      "    accuracy                           1.00    258917\n",
      "   macro avg       1.00      1.00      1.00    258917\n",
      "weighted avg       1.00      1.00      1.00    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                      importance\n",
      "feature                         \n",
      "MI_dir_L3_variance         0.594\n",
      "HH_jit_L1_mean             0.162\n",
      "H_L0.01_mean               0.080\n",
      "HH_L0.1_weight             0.054\n",
      "MI_dir_L0.1_weight         0.053\n",
      "HH_L0.01_magnitude         0.027\n",
      "HH_jit_L0.1_mean           0.010\n",
      "MI_dir_L0.01_mean          0.006\n",
      "HH_jit_L1_weight           0.006\n",
      "H_L0.01_weight             0.003\n",
      "H_L0.1_mean                0.002\n",
      "HH_jit_L0.1_variance       0.001\n",
      "MI_dir_L5_weight           0.000\n",
      "HpHp_L5_weight             0.000\n",
      "HpHp_L5_mean               0.000\n",
      "HpHp_L5_std                0.000\n",
      "HH_jit_L0.01_mean          0.000\n",
      "HpHp_L5_magnitude          0.000\n",
      "HpHp_L5_radius             0.000\n",
      "HpHp_L5_covariance         0.000\n",
      "Random Forest: Accuracy=0.997\n",
      "Random Forest: f1-score=0.997\n",
      "Random Forest: Confusion Matrix [[  5892     14      0]\n",
      " [     0  93785    726]\n",
      " [     4    144 154683]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5906\n",
      "      gafgyt       1.00      0.99      1.00     94511\n",
      "       mirai       1.00      1.00      1.00    154831\n",
      "\n",
      "    accuracy                           1.00    255248\n",
      "   macro avg       1.00      1.00      1.00    255248\n",
      "weighted avg       1.00      1.00      1.00    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with scaling and unbalanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier(dataframe[k], k, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.432\n",
      "H_L0.01_mean                0.185\n",
      "MI_dir_L0.1_mean            0.131\n",
      "HH_jit_L1_mean              0.128\n",
      "H_L0.1_mean                 0.124\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "Random Forest: Accuracy=0.893\n",
      "Random Forest: f1-score=0.865\n",
      "Random Forest: Confusion Matrix [[    1 14995     0     0     0     0     0     1     0     0     6]\n",
      " [    0 18064     0     0     0     0     0     0     0     0     0]\n",
      " [    0  8707     0     0     0     0     0     0     0     0     0]\n",
      " [    0  8827     0     0     0     0     0     0     0     0     0]\n",
      " [    0    33     0     0 27510    29     0     0     0     0     0]\n",
      " [    0    29     0     0     0 31660     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 30707     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 32314     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 36919     0     0]\n",
      " [    0     0     0     0     2     0     0     0     0 71302     0]\n",
      " [    0     0     0     0     0     0     1     0     0     0 24383]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      0.00      0.00     15003\n",
      "  gafgyt_combo       0.36      1.00      0.53     18064\n",
      "   gafgyt_junk       0.00      0.00      0.00      8707\n",
      "   gafgyt_scan       0.00      0.00      0.00      8827\n",
      "    gafgyt_tcp       1.00      1.00      1.00     27572\n",
      "    gafgyt_udp       1.00      1.00      1.00     31689\n",
      "     mirai_ack       1.00      1.00      1.00     30707\n",
      "    mirai_scan       1.00      1.00      1.00     32314\n",
      "     mirai_syn       1.00      1.00      1.00     36919\n",
      "     mirai_udp       1.00      1.00      1.00     71304\n",
      "mirai_udpplain       1.00      1.00      1.00     24384\n",
      "\n",
      "      accuracy                           0.89    305490\n",
      "     macro avg       0.76      0.73      0.68    305490\n",
      "  weighted avg       0.90      0.89      0.87    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.410\n",
      "H_L0.01_mean                0.310\n",
      "HH_jit_L3_mean              0.143\n",
      "MI_dir_L0.01_weight         0.135\n",
      "MI_dir_L0.01_variance       0.001\n",
      "H_L0.1_weight               0.001\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "Random Forest: Accuracy=0.945\n",
      "Random Forest: f1-score=0.925\n",
      "Random Forest: Confusion Matrix [[    1     0     0  4015     0     0     0     0     0     0     0]\n",
      " [    0 15940     0     2     0     0     0     0     0     0     0]\n",
      " [    0  9083     0     3     1     0     0     0     0     0     0]\n",
      " [    0     0     0  8286     0     0     0     0     0     0     0]\n",
      " [    0     0     0    23 27808   492     0     0     0     0     0]\n",
      " [    0     0     0    15     0 31152     0     1     0     0     0]\n",
      " [    0     0     0     0     0     0 34041     0     0     0     0]\n",
      " [    0     1     0     3     0     0     0 13052     1     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35094     0     0]\n",
      " [    0     0     0    54     0     0     0     5     0 45496     0]\n",
      " [    0     0     0     1     0     0     0     1     0     3 26189]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      0.00      0.00      4016\n",
      "  gafgyt_combo       0.64      1.00      0.78     15942\n",
      "   gafgyt_junk       0.00      0.00      0.00      9087\n",
      "   gafgyt_scan       0.67      1.00      0.80      8286\n",
      "    gafgyt_tcp       1.00      0.98      0.99     28323\n",
      "    gafgyt_udp       0.98      1.00      0.99     31168\n",
      "     mirai_ack       1.00      1.00      1.00     34041\n",
      "    mirai_scan       1.00      1.00      1.00     13057\n",
      "     mirai_syn       1.00      1.00      1.00     35094\n",
      "     mirai_udp       1.00      1.00      1.00     45555\n",
      "mirai_udpplain       1.00      1.00      1.00     26194\n",
      "\n",
      "      accuracy                           0.95    250763\n",
      "     macro avg       0.84      0.82      0.78    250763\n",
      "  weighted avg       0.93      0.95      0.92    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.356\n",
      "HH_jit_L1_mean              0.306\n",
      "H_L0.1_weight               0.158\n",
      "MI_dir_L0.01_mean           0.130\n",
      "MI_dir_L0.01_variance       0.048\n",
      "HH_jit_L0.01_mean           0.001\n",
      "H_L0.01_variance            0.001\n",
      "HpHp_L5_std                 0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "Random Forest: Accuracy=0.946\n",
      "Random Forest: f1-score=0.945\n",
      "Random Forest: Confusion Matrix [[11903     0     0     0     0     0]\n",
      " [    2 14183  1839     0     0     0]\n",
      " [    0  3117  5821     0     0     0]\n",
      " [    8     0     0  8512     0     0]\n",
      " [    3     0     0     0 29472   799]\n",
      " [   21     0     0     0     3 30967]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "gafgyt_combo       0.82      0.89      0.85     16024\n",
      " gafgyt_junk       0.76      0.65      0.70      8938\n",
      " gafgyt_scan       1.00      1.00      1.00      8520\n",
      "  gafgyt_tcp       1.00      0.97      0.99     30274\n",
      "  gafgyt_udp       0.97      1.00      0.99     30991\n",
      "\n",
      "    accuracy                           0.95    106650\n",
      "   macro avg       0.93      0.92      0.92    106650\n",
      "weighted avg       0.95      0.95      0.94    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.353\n",
      "MI_dir_L0.01_mean           0.352\n",
      "HpHp_L0.01_radius           0.177\n",
      "HH_jit_L1_mean              0.114\n",
      "H_L0.01_variance            0.003\n",
      "HH_L0.01_covariance         0.000\n",
      "HH_L0.01_pcc                0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_L0.01_magnitude          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "Random Forest: Accuracy=0.884\n",
      "Random Forest: f1-score=0.857\n",
      "Random Forest: Confusion Matrix [[49673     0     0     0     0     0     0     0   135     0  2690]\n",
      " [    0     5     2     0     0     0     0     0     0     0 17511]\n",
      " [    0     0     2     0     0     0     0     0     0     0  8475]\n",
      " [    0     0     0    22     0     0     0     0     0     0  8208]\n",
      " [    0     0     0     0 26880   999     0     0     0     0    26]\n",
      " [    6     0     0     0     0 31716     0     0     0     0    28]\n",
      " [    0     0     0     0     0     0 27317     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 31096     1     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35281     0     7]\n",
      " [    0     0     0     0     1     0     0     0     0 65060     0]\n",
      " [    0     0     0     0     0     0     1     0     0     0 24462]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      0.95      0.97     52498\n",
      "  gafgyt_combo       1.00      0.00      0.00     17518\n",
      "   gafgyt_junk       0.50      0.00      0.00      8477\n",
      "   gafgyt_scan       1.00      0.00      0.01      8230\n",
      "    gafgyt_tcp       1.00      0.96      0.98     27905\n",
      "    gafgyt_udp       0.97      1.00      0.98     31750\n",
      "     mirai_ack       1.00      1.00      1.00     27317\n",
      "    mirai_scan       1.00      1.00      1.00     31097\n",
      "     mirai_syn       1.00      1.00      1.00     35288\n",
      "     mirai_udp       1.00      1.00      1.00     65061\n",
      "mirai_udpplain       0.40      1.00      0.57     24463\n",
      "\n",
      "      accuracy                           0.88    329604\n",
      "     macro avg       0.90      0.72      0.68    329604\n",
      "  weighted avg       0.94      0.88      0.86    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.486\n",
      "HH_jit_L1_mean              0.166\n",
      "MI_dir_L0.01_weight         0.163\n",
      "MI_dir_L0.1_weight          0.104\n",
      "H_L0.01_variance            0.079\n",
      "H_L1_weight                 0.001\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "Random Forest: Accuracy=0.810\n",
      "Random Forest: f1-score=0.781\n",
      "Random Forest: Confusion Matrix [[18758     0     0     0     0     0     0     0     0     0     0]\n",
      " [    1 18405     0     0     0     0     0     0     0     0     0]\n",
      " [    4  9173     0     0     0     0     0     0     0     0     0]\n",
      " [    1  8619     0     0     1     0     0     0     0     0     0]\n",
      " [   30     0     0     0 29301  1808     0     0     0     0     0]\n",
      " [   30     0     0     0     1 31232     0     0     0     0     0]\n",
      " [    0 10469     0     0     0     0  7696     0     0     0     0]\n",
      " [    0     1     0     0     0     0     0 28964     0     0     0]\n",
      " [    4     0     0     0     0     0     0     0 19837     2     0]\n",
      " [    1     9     0     0     0     0     0     0     0 46971     0]\n",
      " [    0 17160     0     0     0     0     0     0     0     0     0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     18758\n",
      "  gafgyt_combo       0.29      1.00      0.45     18406\n",
      "   gafgyt_junk       0.00      0.00      0.00      9177\n",
      "   gafgyt_scan       0.00      0.00      0.00      8621\n",
      "    gafgyt_tcp       1.00      0.94      0.97     31139\n",
      "    gafgyt_udp       0.95      1.00      0.97     31263\n",
      "     mirai_ack       1.00      0.42      0.60     18165\n",
      "    mirai_scan       1.00      1.00      1.00     28965\n",
      "     mirai_syn       1.00      1.00      1.00     19843\n",
      "     mirai_udp       1.00      1.00      1.00     46981\n",
      "mirai_udpplain       0.00      0.00      0.00     17160\n",
      "\n",
      "      accuracy                           0.81    248478\n",
      "     macro avg       0.66      0.67      0.63    248478\n",
      "  weighted avg       0.80      0.81      0.78    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.313\n",
      "H_L0.1_mean                 0.296\n",
      "HH_jit_L1_mean              0.165\n",
      "H_L0.1_weight               0.154\n",
      "MI_dir_L0.01_variance       0.070\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "Random Forest: Accuracy=0.825\n",
      "Random Forest: f1-score=0.799\n",
      "Random Forest: Confusion Matrix [[29696     0     0     0     0     0     0     0     0     2     0]\n",
      " [    5 17164     0     0     0     0     0     0     0     0     0]\n",
      " [    4  8588     0     0     0     0     0     0     0     0     0]\n",
      " [    3  8560     0     0     0     0     0     0     0     0     0]\n",
      " [    4     0     0     0 26569     5     0     0     0     0     0]\n",
      " [   15     0     0     0     3 31327     0     0     0     0     0]\n",
      " [    0 10450     0     0     0     0  7027     0     0     0     0]\n",
      " [    2     0     0     0     0     0     0 29204     0     0     0]\n",
      " [    1     0     0     0     0     0     0     0 18635     0     0]\n",
      " [    0     8     0     0     1     0     0     0     0 47539     0]\n",
      " [    0 16255     0     0     1     0     0     0     0     0     0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     29698\n",
      "  gafgyt_combo       0.28      1.00      0.44     17169\n",
      "   gafgyt_junk       0.00      0.00      0.00      8592\n",
      "   gafgyt_scan       0.00      0.00      0.00      8563\n",
      "    gafgyt_tcp       1.00      1.00      1.00     26578\n",
      "    gafgyt_udp       1.00      1.00      1.00     31345\n",
      "     mirai_ack       1.00      0.40      0.57     17477\n",
      "    mirai_scan       1.00      1.00      1.00     29206\n",
      "     mirai_syn       1.00      1.00      1.00     18636\n",
      "     mirai_udp       1.00      1.00      1.00     47548\n",
      "mirai_udpplain       0.00      0.00      0.00     16256\n",
      "\n",
      "      accuracy                           0.83    251068\n",
      "     macro avg       0.66      0.67      0.64    251068\n",
      "  weighted avg       0.82      0.83      0.80    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.339\n",
      "HH_jit_L1_mean              0.317\n",
      "H_L0.01_mean                0.184\n",
      "MI_dir_L0.01_mean           0.123\n",
      "HH_jit_L0.1_mean            0.036\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.922\n",
      "Random Forest: f1-score=0.892\n",
      "Random Forest: Confusion Matrix [[15675     5     0     0     1     3]\n",
      " [    1 17688     0     0     0     1]\n",
      " [    0  8604     1     0     0     1]\n",
      " [    1     0     0  8194     0     2]\n",
      " [   17     1     0     0 29240    79]\n",
      " [   20     1     0     0     1 33031]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "gafgyt_combo       0.67      1.00      0.80     17690\n",
      " gafgyt_junk       1.00      0.00      0.00      8606\n",
      " gafgyt_scan       1.00      1.00      1.00      8197\n",
      "  gafgyt_tcp       1.00      1.00      1.00     29337\n",
      "  gafgyt_udp       1.00      1.00      1.00     33053\n",
      "\n",
      "    accuracy                           0.92    112567\n",
      "   macro avg       0.94      0.83      0.80    112567\n",
      "weighted avg       0.95      0.92      0.89    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.732\n",
      "HH_jit_L0.1_mean            0.138\n",
      "H_L0.01_weight              0.130\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "Random Forest: Accuracy=0.935\n",
      "Random Forest: f1-score=0.913\n",
      "Random Forest: Confusion Matrix [[14050     5     0     0     0     0     0     0     0     0     0]\n",
      " [    2 16316     0     0     0     0     0     0     0     0     0]\n",
      " [    1  8622     0     0     0     0     0     0     0     0     0]\n",
      " [    3  8209     0     0     1     0     0     0     0     0     0]\n",
      " [   21     0     0     0 26427     2     0     0     0     0     0]\n",
      " [   36     1     0     0     0 30892     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 33471     0     0     0     0]\n",
      " [    3     2     0     0     0     0     0 13872     1     0     0]\n",
      " [    0     0     0     0     0     0     0     0 37909     0     0]\n",
      " [    0     0     0     0     0     0     0     0     1 45602     0]\n",
      " [    0     0     0     0     0     0     0     0     0     1 23467]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     14055\n",
      "  gafgyt_combo       0.49      1.00      0.66     16318\n",
      "   gafgyt_junk       0.00      0.00      0.00      8623\n",
      "   gafgyt_scan       0.00      0.00      0.00      8213\n",
      "    gafgyt_tcp       1.00      1.00      1.00     26450\n",
      "    gafgyt_udp       1.00      1.00      1.00     30929\n",
      "     mirai_ack       1.00      1.00      1.00     33471\n",
      "    mirai_scan       1.00      1.00      1.00     13878\n",
      "     mirai_syn       1.00      1.00      1.00     37909\n",
      "     mirai_udp       1.00      1.00      1.00     45603\n",
      "mirai_udpplain       1.00      1.00      1.00     23468\n",
      "\n",
      "      accuracy                           0.93    258917\n",
      "     macro avg       0.77      0.82      0.79    258917\n",
      "  weighted avg       0.90      0.93      0.91    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.399\n",
      "MI_dir_L0.01_mean           0.188\n",
      "H_L0.01_weight              0.136\n",
      "HH_jit_L1_mean              0.131\n",
      "H_L0.1_mean                 0.095\n",
      "MI_dir_L0.1_weight          0.049\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L3_weight              0.000\n",
      "Random Forest: Accuracy=0.937\n",
      "Random Forest: f1-score=0.915\n",
      "Random Forest: Confusion Matrix [[   16     0     0  5890     0     0     0     0     0     0     0]\n",
      " [    0 17827     0     5     0     0     0     0     0     0     0]\n",
      " [    0  8264     0     2     0     0     0     0     0     0     0]\n",
      " [    0     0     0  8570     0     0     0     0     0     0     0]\n",
      " [    0     0     0    29 27304  1835     0     0     0     0     0]\n",
      " [    0     0     0    26     0 30649     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 32254     0     0     0     0]\n",
      " [    0     1     0     5     0     0     0 13081     0     0     0]\n",
      " [    0     0     0     0     0     0     0     1 36878     0     0]\n",
      " [    0     0     0     0     0     0     0     1     0 47133     0]\n",
      " [    0     0     0     1     0     0     3     0     0     0 25473]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      0.00      0.01      5906\n",
      "  gafgyt_combo       0.68      1.00      0.81     17832\n",
      "   gafgyt_junk       0.00      0.00      0.00      8266\n",
      "   gafgyt_scan       0.59      1.00      0.74      8570\n",
      "    gafgyt_tcp       1.00      0.94      0.97     29168\n",
      "    gafgyt_udp       0.94      1.00      0.97     30675\n",
      "     mirai_ack       1.00      1.00      1.00     32254\n",
      "    mirai_scan       1.00      1.00      1.00     13087\n",
      "     mirai_syn       1.00      1.00      1.00     36879\n",
      "     mirai_udp       1.00      1.00      1.00     47134\n",
      "mirai_udpplain       1.00      1.00      1.00     25477\n",
      "\n",
      "      accuracy                           0.94    255248\n",
      "     macro avg       0.84      0.81      0.77    255248\n",
      "  weighted avg       0.92      0.94      0.92    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# without scaling and unbalanced data multi label\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier(dataframe_multi[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.603\n",
      "H_L0.01_mean                0.255\n",
      "H_L0.1_mean                 0.142\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.803\n",
      "Random Forest: f1-score=0.744\n",
      "Random Forest: Confusion Matrix [[    1 14995     0     0     0     0     0     1     0     0     6]\n",
      " [    0 18064     0     0     0     0     0     0     0     0     0]\n",
      " [    0  8707     0     0     0     0     0     0     0     0     0]\n",
      " [    0  8827     0     0     0     0     0     0     0     0     0]\n",
      " [    0    33     0     0     0 27539     0     0     0     0     0]\n",
      " [    0    29     0     0     0 31660     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 30707     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 32314     0     0     0]\n",
      " [    0     0     0     0     0     0     0     0 36919     0     0]\n",
      " [    0     0     0     0     0     0     0     0     0 71304     0]\n",
      " [    0     0     0     0     0     0     1     0     0     0 24383]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      0.00      0.00     15003\n",
      "  gafgyt_combo       0.36      1.00      0.53     18064\n",
      "   gafgyt_junk       0.00      0.00      0.00      8707\n",
      "   gafgyt_scan       0.00      0.00      0.00      8827\n",
      "    gafgyt_tcp       0.00      0.00      0.00     27572\n",
      "    gafgyt_udp       0.53      1.00      0.70     31689\n",
      "     mirai_ack       1.00      1.00      1.00     30707\n",
      "    mirai_scan       1.00      1.00      1.00     32314\n",
      "     mirai_syn       1.00      1.00      1.00     36919\n",
      "     mirai_udp       1.00      1.00      1.00     71304\n",
      "mirai_udpplain       1.00      1.00      1.00     24384\n",
      "\n",
      "      accuracy                           0.80    305490\n",
      "     macro avg       0.63      0.64      0.57    305490\n",
      "  weighted avg       0.77      0.80      0.74    305490\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.584\n",
      "MI_dir_L0.01_mean           0.256\n",
      "MI_dir_L0.01_weight         0.158\n",
      "H_L0.1_variance             0.001\n",
      "MI_dir_L0.1_weight          0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "Random Forest: Accuracy=0.834\n",
      "Random Forest: f1-score=0.775\n",
      "Random Forest: Confusion Matrix [[    6     0     0  4010     0     0     0     0     0     0     0]\n",
      " [    0 15940     0     2     0     0     0     0     0     0     0]\n",
      " [    0  9083     0     3     0     1     0     0     0     0     0]\n",
      " [    0     0     0  8286     0     0     0     0     0     0     0]\n",
      " [    0     0     0    23     0 28300     0     0     0     0     0]\n",
      " [    0     0     0    15     0 31152     0     1     0     0     0]\n",
      " [    0     0     0     0     0     0 34038     0     0     3     0]\n",
      " [    0     1     0     3     0     0     0 13052     1     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35094     0     0]\n",
      " [    0     0     0    54     0     0     0     5     0 45496     0]\n",
      " [    0     0     0     1     0     0     0     1     0     4 26188]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      0.00      0.00      4016\n",
      "  gafgyt_combo       0.64      1.00      0.78     15942\n",
      "   gafgyt_junk       0.00      0.00      0.00      9087\n",
      "   gafgyt_scan       0.67      1.00      0.80      8286\n",
      "    gafgyt_tcp       0.00      0.00      0.00     28323\n",
      "    gafgyt_udp       0.52      1.00      0.69     31168\n",
      "     mirai_ack       1.00      1.00      1.00     34041\n",
      "    mirai_scan       1.00      1.00      1.00     13057\n",
      "     mirai_syn       1.00      1.00      1.00     35094\n",
      "     mirai_udp       1.00      1.00      1.00     45555\n",
      "mirai_udpplain       1.00      1.00      1.00     26194\n",
      "\n",
      "      accuracy                           0.83    250763\n",
      "     macro avg       0.71      0.73      0.66    250763\n",
      "  weighted avg       0.76      0.83      0.78    250763\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.624\n",
      "H_L0.1_weight               0.248\n",
      "H_L0.01_variance            0.076\n",
      "H_L0.1_mean                 0.034\n",
      "HH_L0.01_std                0.016\n",
      "H_L0.1_variance             0.001\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_weight              0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "Random Forest: Accuracy=0.670\n",
      "Random Forest: f1-score=0.574\n",
      "Random Forest: Confusion Matrix [[11903     0     0     0     0     0]\n",
      " [    4 14181  1839     0     0     0]\n",
      " [    1  3117  5820     0     0     0]\n",
      " [    8     0     0  8512     0     0]\n",
      " [    5     0     0     0    21 30248]\n",
      " [   19     0     0     0     1 30971]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11903\n",
      "gafgyt_combo       0.82      0.88      0.85     16024\n",
      " gafgyt_junk       0.76      0.65      0.70      8938\n",
      " gafgyt_scan       1.00      1.00      1.00      8520\n",
      "  gafgyt_tcp       0.95      0.00      0.00     30274\n",
      "  gafgyt_udp       0.51      1.00      0.67     30991\n",
      "\n",
      "    accuracy                           0.67    106650\n",
      "   macro avg       0.84      0.76      0.70    106650\n",
      "weighted avg       0.80      0.67      0.57    106650\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.575\n",
      "MI_dir_L0.01_mean           0.220\n",
      "HpHp_L0.01_radius           0.200\n",
      "MI_dir_L0.01_variance       0.004\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "Random Forest: Accuracy=0.803\n",
      "Random Forest: f1-score=0.746\n",
      "Random Forest: Confusion Matrix [[49672     0     0     0     0     1     0     0   135     0  2690]\n",
      " [    0     5     2     0     0     0     0     0     0     0 17511]\n",
      " [    0     0     2     0     0     0     0     0     0     0  8475]\n",
      " [    0     0     0    22     0     0     0     0     0     0  8208]\n",
      " [    0     0     0     0     2 27877     0     0     0     0    26]\n",
      " [    5     0     0     0     0 31717     0     0     0     0    28]\n",
      " [    0     0     0     0     0     0 27317     0     0     0     0]\n",
      " [    0     0     0     0     0     0     0 31096     1     0     0]\n",
      " [    0     0     0     0     0     0     0     0 35281     0     7]\n",
      " [    0     0     0     0     0     1     0     0     0 65060     0]\n",
      " [    0     0     0     0     0     0     1     0     0     0 24462]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      0.95      0.97     52498\n",
      "  gafgyt_combo       1.00      0.00      0.00     17518\n",
      "   gafgyt_junk       0.50      0.00      0.00      8477\n",
      "   gafgyt_scan       1.00      0.00      0.01      8230\n",
      "    gafgyt_tcp       1.00      0.00      0.00     27905\n",
      "    gafgyt_udp       0.53      1.00      0.69     31750\n",
      "     mirai_ack       1.00      1.00      1.00     27317\n",
      "    mirai_scan       1.00      1.00      1.00     31097\n",
      "     mirai_syn       1.00      1.00      1.00     35288\n",
      "     mirai_udp       1.00      1.00      1.00     65061\n",
      "mirai_udpplain       0.40      1.00      0.57     24463\n",
      "\n",
      "      accuracy                           0.80    329604\n",
      "     macro avg       0.86      0.63      0.57    329604\n",
      "  weighted avg       0.90      0.80      0.75    329604\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.262\n",
      "MI_dir_L0.01_weight         0.196\n",
      "MI_dir_L0.1_mean            0.187\n",
      "H_L0.01_mean                0.134\n",
      "H_L0.1_weight               0.125\n",
      "MI_dir_L0.01_variance       0.095\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_weight              0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "Random Forest: Accuracy=0.691\n",
      "Random Forest: f1-score=0.621\n",
      "Random Forest: Confusion Matrix [[18758     0     0     0     0     0     0     0     0     0     0]\n",
      " [    1 18405     0     0     0     0     0     0     0     0     0]\n",
      " [    4  9173     0     0     0     0     0     0     0     0     0]\n",
      " [    1  8619     0     0     1     0     0     0     0     0     0]\n",
      " [   30     0     0     0 31109     0     0     0     0     0     0]\n",
      " [   30     0     0     0 31231     2     0     0     0     0     0]\n",
      " [    0 10469     0     0     0     0  7696     0     0     0     0]\n",
      " [    0     1     0     0     0     0     0 28964     0     0     0]\n",
      " [    4     0     0     0     0     0     0     0 19837     2     0]\n",
      " [    2     9     0     0     0     0     0     0     0 46970     0]\n",
      " [    0 17160     0     0     0     0     0     0     0     0     0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     18758\n",
      "  gafgyt_combo       0.29      1.00      0.45     18406\n",
      "   gafgyt_junk       0.00      0.00      0.00      9177\n",
      "   gafgyt_scan       0.00      0.00      0.00      8621\n",
      "    gafgyt_tcp       0.50      1.00      0.67     31139\n",
      "    gafgyt_udp       1.00      0.00      0.00     31263\n",
      "     mirai_ack       1.00      0.42      0.60     18165\n",
      "    mirai_scan       1.00      1.00      1.00     28965\n",
      "     mirai_syn       1.00      1.00      1.00     19843\n",
      "     mirai_udp       1.00      1.00      1.00     46981\n",
      "mirai_udpplain       0.00      0.00      0.00     17160\n",
      "\n",
      "      accuracy                           0.69    248478\n",
      "     macro avg       0.62      0.58      0.52    248478\n",
      "  weighted avg       0.74      0.69      0.62    248478\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_mean                 0.355\n",
      "MI_dir_L0.01_mean           0.257\n",
      "H_L0.1_weight               0.185\n",
      "MI_dir_L0.1_mean            0.118\n",
      "H_L0.01_variance            0.084\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_weight              0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "Random Forest: Accuracy=0.719\n",
      "Random Forest: f1-score=0.656\n",
      "Random Forest: Confusion Matrix [[29696     0     0     0     0     0     0     0     0     2     0]\n",
      " [    5 17164     0     0     0     0     0     0     0     0     0]\n",
      " [    4  8588     0     0     0     0     0     0     0     0     0]\n",
      " [    3  8560     0     0     0     0     0     0     0     0     0]\n",
      " [    4     0     0     0    16 26558     0     0     0     0     0]\n",
      " [   15     0     0     0     2 31328     0     0     0     0     0]\n",
      " [    0 10450     0     0     0     0  7026     0     0     0     1]\n",
      " [    2     0     0     0     0     0     0 29204     0     0     0]\n",
      " [    1     0     0     0     0     0     0     0 18635     0     0]\n",
      " [    0     8     0     0     1     0     0     0     0 47539     0]\n",
      " [    0 16255     0     0     1     0     0     0     0     0     0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     29698\n",
      "  gafgyt_combo       0.28      1.00      0.44     17169\n",
      "   gafgyt_junk       0.00      0.00      0.00      8592\n",
      "   gafgyt_scan       0.00      0.00      0.00      8563\n",
      "    gafgyt_tcp       0.80      0.00      0.00     26578\n",
      "    gafgyt_udp       0.54      1.00      0.70     31345\n",
      "     mirai_ack       1.00      0.40      0.57     17477\n",
      "    mirai_scan       1.00      1.00      1.00     29206\n",
      "     mirai_syn       1.00      1.00      1.00     18636\n",
      "     mirai_udp       1.00      1.00      1.00     47548\n",
      "mirai_udpplain       0.00      0.00      0.00     16256\n",
      "\n",
      "      accuracy                           0.72    251068\n",
      "     macro avg       0.60      0.58      0.52    251068\n",
      "  weighted avg       0.74      0.72      0.66    251068\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_mean                0.479\n",
      "MI_dir_L0.01_mean           0.420\n",
      "H_L0.01_weight              0.055\n",
      "H_L0.01_variance            0.044\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_mean                0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "Random Forest: Accuracy=0.701\n",
      "Random Forest: f1-score=0.611\n",
      "Random Forest: Confusion Matrix [[15678     1     3     0     0     2]\n",
      " [    2 14764  2924     0     0     0]\n",
      " [    0  1381  7225     0     0     0]\n",
      " [    1     2     0  8194     0     0]\n",
      " [   18     0     0     0    10 29309]\n",
      " [   21     1     0     0     0 33031]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15684\n",
      "gafgyt_combo       0.91      0.83      0.87     17690\n",
      " gafgyt_junk       0.71      0.84      0.77      8606\n",
      " gafgyt_scan       1.00      1.00      1.00      8197\n",
      "  gafgyt_tcp       1.00      0.00      0.00     29337\n",
      "  gafgyt_udp       0.53      1.00      0.69     33053\n",
      "\n",
      "    accuracy                           0.70    112567\n",
      "   macro avg       0.86      0.78      0.72    112567\n",
      "weighted avg       0.83      0.70      0.61    112567\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.523\n",
      "H_L0.01_mean                0.327\n",
      "H_L0.01_weight              0.150\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.833\n",
      "Random Forest: f1-score=0.775\n",
      "Random Forest: Confusion Matrix [[14044     5     0     0     0     6     0     0     0     0     0]\n",
      " [    1 16316     0     0     0     1     0     0     0     0     0]\n",
      " [    0  8622     0     0     0     1     0     0     0     0     0]\n",
      " [    3  8209     0     0     0     1     0     0     0     0     0]\n",
      " [   20     0     0     0     0 26430     0     0     0     0     0]\n",
      " [   36     1     0     0     0 30892     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 33471     0     0     0     0]\n",
      " [    2     2     0     0     0     0     0 13872     2     0     0]\n",
      " [    0     0     0     0     0     0     0     0 37909     0     0]\n",
      " [    0     0     0     0     0     0     0     0     1 45602     0]\n",
      " [    0     0     0     0     0     0     0     0     0     1 23467]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     14055\n",
      "  gafgyt_combo       0.49      1.00      0.66     16318\n",
      "   gafgyt_junk       0.00      0.00      0.00      8623\n",
      "   gafgyt_scan       0.00      0.00      0.00      8213\n",
      "    gafgyt_tcp       0.00      0.00      0.00     26450\n",
      "    gafgyt_udp       0.54      1.00      0.70     30929\n",
      "     mirai_ack       1.00      1.00      1.00     33471\n",
      "    mirai_scan       1.00      1.00      1.00     13878\n",
      "     mirai_syn       1.00      1.00      1.00     37909\n",
      "     mirai_udp       1.00      1.00      1.00     45603\n",
      "mirai_udpplain       1.00      1.00      1.00     23468\n",
      "\n",
      "      accuracy                           0.83    258917\n",
      "     macro avg       0.64      0.73      0.67    258917\n",
      "  weighted avg       0.75      0.83      0.78    258917\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.471\n",
      "H_L0.01_mean                0.205\n",
      "H_L0.01_weight              0.157\n",
      "MI_dir_L0.1_mean            0.109\n",
      "H_L0.1_weight               0.057\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_weight              0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "Random Forest: Accuracy=0.830\n",
      "Random Forest: f1-score=0.770\n",
      "Random Forest: Confusion Matrix [[   16     0     0  5890     0     0     0     0     0     0     0]\n",
      " [    1 17827     0     4     0     0     0     0     0     0     0]\n",
      " [    0  8264     0     2     0     0     0     0     0     0     0]\n",
      " [    0     0     0  8570     0     0     0     0     0     0     0]\n",
      " [    0     0     0    29     0 29139     0     0     0     0     0]\n",
      " [    0     0     0    26     0 30649     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 32254     0     0     0     0]\n",
      " [    2     1     0     5     0     0     0 13079     0     0     0]\n",
      " [    0     0     0     0     0     0     0     1 36878     0     0]\n",
      " [    1     0     0     0     0     0     0     0     0 47133     0]\n",
      " [    0     0     0     1     0     0     0     0     0     3 25473]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       0.80      0.00      0.01      5906\n",
      "  gafgyt_combo       0.68      1.00      0.81     17832\n",
      "   gafgyt_junk       0.00      0.00      0.00      8266\n",
      "   gafgyt_scan       0.59      1.00      0.74      8570\n",
      "    gafgyt_tcp       0.00      0.00      0.00     29168\n",
      "    gafgyt_udp       0.51      1.00      0.68     30675\n",
      "     mirai_ack       1.00      1.00      1.00     32254\n",
      "    mirai_scan       1.00      1.00      1.00     13087\n",
      "     mirai_syn       1.00      1.00      1.00     36879\n",
      "     mirai_udp       1.00      1.00      1.00     47134\n",
      "mirai_udpplain       1.00      1.00      1.00     25477\n",
      "\n",
      "      accuracy                           0.83    255248\n",
      "     macro avg       0.69      0.73      0.66    255248\n",
      "  weighted avg       0.75      0.83      0.77    255248\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# without scaling and unbalanced data multi label\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier(dataframe_multi[k], k, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dt_classifier_balanced(data, device_name, scaling=False):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "    nm = NearMiss()\n",
    "    \n",
    "    X_res, y_res = nm.fit_resample(X, y)\n",
    "    print(f'Imblearn Shape:{X_res.shape}, {y_res.shape}')\n",
    "    \n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.30, random_state=47)\n",
    "        model_name = f'{device_name}_without_scaling_balanced_model_decision_tree_gini.pkl'\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        model_name = f'{device_name}_with_scaling_balanced_model_decision_tree_gini.pkl'\n",
    "    \n",
    "    clf = DecisionTreeClassifier(criterion='gini', random_state= 1234, max_depth  = 5 )\n",
    "    model_res = clf.fit(X_train, y_train)\n",
    "    y_pred = model_res.predict(X_test)\n",
    "    y_pred_prob = model_res.predict_proba(X_test)\n",
    "    lr_probs = y_pred_prob[:,1]\n",
    "    ac = accuracy_score(y_test, y_pred)\n",
    "    importances = pd.DataFrame({'feature':X.columns,'importance':np.round(clf.feature_importances_,3)})\n",
    "    importances = importances.sort_values('importance',ascending=False).set_index('feature')\n",
    "\n",
    "    f1 = f1_score(y_test, y_pred, average='weighted')\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "    print(importances.head(20))\n",
    "    \n",
    "    print('Random Forest: Accuracy=%.3f' % (ac))\n",
    "\n",
    "    print('Random Forest: f1-score=%.3f' % (f1))\n",
    "    print('Random Forest: Confusion Matrix', cm)\n",
    "    print('Random Forest : Classification Report',classification_report(y_test, y_pred))\n",
    "    \n",
    "    best_model = clf\n",
    "    best_model.version = 1.0\n",
    "    best_model.pandas_version = pd.__version__\n",
    "    best_model.numpy_version = np.__version__\n",
    "    best_model.sklearn_version = sklearn_version\n",
    "    #best_model.X_columns = [col for col in X_train.columns]\n",
    "    best_model.build_datetime = datetime.now()\n",
    "    \n",
    "    modelpath = 'models'\n",
    "    if not os.path.exists(modelpath):\n",
    "        os.mkdir(modelpath)\n",
    "    iotmodel_path = os.path.join(modelpath, model_name)\n",
    "    if not os.path.exists(iotmodel_path):\n",
    "        with open(iotmodel_path, 'wb') as f:\n",
    "            pickle.dump(best_model, f)\n",
    "    \n",
    "    print('---------------------- Done ---------------------')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(148644, 115), (148644,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean                0.5\n",
      "HH_jit_L0.1_mean              0.5\n",
      "MI_dir_L5_weight              0.0\n",
      "HH_jit_L1_variance            0.0\n",
      "HpHp_L5_radius                0.0\n",
      "HpHp_L5_magnitude             0.0\n",
      "HpHp_L5_std                   0.0\n",
      "HpHp_L5_mean                  0.0\n",
      "HpHp_L5_weight                0.0\n",
      "HH_jit_L0.01_variance         0.0\n",
      "HH_jit_L0.01_mean             0.0\n",
      "HH_jit_L0.01_weight           0.0\n",
      "HH_jit_L0.1_variance          0.0\n",
      "HH_jit_L0.1_weight            0.0\n",
      "HH_jit_L1_mean                0.0\n",
      "HpHp_L5_pcc                   0.0\n",
      "HH_jit_L1_weight              0.0\n",
      "HH_jit_L3_variance            0.0\n",
      "HH_jit_L3_mean                0.0\n",
      "HH_jit_L3_weight              0.0\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[14849     0     0]\n",
      " [    0 14860     0]\n",
      " [    0     0 14885]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14849\n",
      "      gafgyt       1.00      1.00      1.00     14860\n",
      "       mirai       1.00      1.00      1.00     14885\n",
      "\n",
      "    accuracy                           1.00     44594\n",
      "   macro avg       1.00      1.00      1.00     44594\n",
      "weighted avg       1.00      1.00      1.00     44594\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(39339, 115), (39339,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.501\n",
      "HH_jit_L3_mean              0.499\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[3913    0    0]\n",
      " [   0 3891    0]\n",
      " [   0    0 3998]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      3913\n",
      "      gafgyt       1.00      1.00      1.00      3891\n",
      "       mirai       1.00      1.00      1.00      3998\n",
      "\n",
      "    accuracy                           1.00     11802\n",
      "   macro avg       1.00      1.00      1.00     11802\n",
      "weighted avg       1.00      1.00      1.00     11802\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(78200, 115), (78200,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.997\n",
      "HH_jit_L3_mean              0.003\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[11788     0]\n",
      " [    0 11672]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11788\n",
      "      gafgyt       1.00      1.00      1.00     11672\n",
      "\n",
      "    accuracy                           1.00     23460\n",
      "   macro avg       1.00      1.00      1.00     23460\n",
      "weighted avg       1.00      1.00      1.00     23460\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(525720, 115), (525720,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.500\n",
      "HH_jit_L0.1_mean            0.499\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[52311     0     0]\n",
      " [    0 52769     0]\n",
      " [    0     0 52636]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52311\n",
      "      gafgyt       1.00      1.00      1.00     52769\n",
      "       mirai       1.00      1.00      1.00     52636\n",
      "\n",
      "    accuracy                           1.00    157716\n",
      "   macro avg       1.00      1.00      1.00    157716\n",
      "weighted avg       1.00      1.00      1.00    157716\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(186462, 115), (186462,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean                0.5\n",
      "HH_jit_L0.1_mean              0.5\n",
      "MI_dir_L5_weight              0.0\n",
      "HH_jit_L1_variance            0.0\n",
      "HpHp_L5_radius                0.0\n",
      "HpHp_L5_magnitude             0.0\n",
      "HpHp_L5_std                   0.0\n",
      "HpHp_L5_mean                  0.0\n",
      "HpHp_L5_weight                0.0\n",
      "HH_jit_L0.01_variance         0.0\n",
      "HH_jit_L0.01_mean             0.0\n",
      "HH_jit_L0.01_weight           0.0\n",
      "HH_jit_L0.1_variance          0.0\n",
      "HH_jit_L0.1_weight            0.0\n",
      "HH_jit_L1_mean                0.0\n",
      "HpHp_L5_pcc                   0.0\n",
      "HH_jit_L1_weight              0.0\n",
      "HH_jit_L3_variance            0.0\n",
      "HH_jit_L3_mean                0.0\n",
      "HH_jit_L3_weight              0.0\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[18587     0     0]\n",
      " [    0 18646     0]\n",
      " [    0     0 18706]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18587\n",
      "      gafgyt       1.00      1.00      1.00     18646\n",
      "       mirai       1.00      1.00      1.00     18706\n",
      "\n",
      "    accuracy                           1.00     55939\n",
      "   macro avg       1.00      1.00      1.00     55939\n",
      "weighted avg       1.00      1.00      1.00     55939\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(295542, 115), (295542,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.500\n",
      "HH_jit_L1_mean              0.499\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[29478     0     0]\n",
      " [    0 29701     0]\n",
      " [    0     0 29484]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29478\n",
      "      gafgyt       1.00      1.00      1.00     29701\n",
      "       mirai       1.00      1.00      1.00     29484\n",
      "\n",
      "    accuracy                           1.00     88663\n",
      "   macro avg       1.00      1.00      1.00     88663\n",
      "weighted avg       1.00      1.00      1.00     88663\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(104300, 115), (104300,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.999\n",
      "HH_jit_L5_mean              0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15642     0]\n",
      " [    0 15648]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15642\n",
      "      gafgyt       1.00      1.00      1.00     15648\n",
      "\n",
      "    accuracy                           1.00     31290\n",
      "   macro avg       1.00      1.00      1.00     31290\n",
      "weighted avg       1.00      1.00      1.00     31290\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(139755, 115), (139755,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean                0.5\n",
      "HH_jit_L5_mean                0.5\n",
      "HH_jit_L1_variance            0.0\n",
      "HpHp_L5_radius                0.0\n",
      "HpHp_L5_magnitude             0.0\n",
      "HpHp_L5_std                   0.0\n",
      "HpHp_L5_mean                  0.0\n",
      "HpHp_L5_weight                0.0\n",
      "HH_jit_L0.01_variance         0.0\n",
      "HH_jit_L0.01_mean             0.0\n",
      "HH_jit_L0.01_weight           0.0\n",
      "HH_jit_L0.1_variance          0.0\n",
      "HH_jit_L0.1_mean              0.0\n",
      "HH_jit_L0.1_weight            0.0\n",
      "MI_dir_L5_weight              0.0\n",
      "HpHp_L5_pcc                   0.0\n",
      "HH_jit_L1_weight              0.0\n",
      "HH_jit_L3_variance            0.0\n",
      "HH_jit_L3_mean                0.0\n",
      "HH_jit_L3_weight              0.0\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[13999     0     0]\n",
      " [    0 13971     0]\n",
      " [    0     0 13957]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     13999\n",
      "      gafgyt       1.00      1.00      1.00     13971\n",
      "       mirai       1.00      1.00      1.00     13957\n",
      "\n",
      "    accuracy                           1.00     41927\n",
      "   macro avg       1.00      1.00      1.00     41927\n",
      "weighted avg       1.00      1.00      1.00     41927\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(58584, 115), (58584,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.502\n",
      "HH_jit_L0.1_mean            0.498\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[5743    0    0]\n",
      " [   0 5859    0]\n",
      " [   0    0 5974]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5743\n",
      "      gafgyt       1.00      1.00      1.00      5859\n",
      "       mirai       1.00      1.00      1.00      5974\n",
      "\n",
      "    accuracy                           1.00     17576\n",
      "   macro avg       1.00      1.00      1.00     17576\n",
      "weighted avg       1.00      1.00      1.00     17576\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with out scaling and unbalanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier_balanced(dataframe[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "Danmin Door Bell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(148644, 115), (148644,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean                0.5\n",
      "HH_jit_L0.1_mean              0.5\n",
      "MI_dir_L5_weight              0.0\n",
      "HH_jit_L1_variance            0.0\n",
      "HpHp_L5_radius                0.0\n",
      "HpHp_L5_magnitude             0.0\n",
      "HpHp_L5_std                   0.0\n",
      "HpHp_L5_mean                  0.0\n",
      "HpHp_L5_weight                0.0\n",
      "HH_jit_L0.01_variance         0.0\n",
      "HH_jit_L0.01_mean             0.0\n",
      "HH_jit_L0.01_weight           0.0\n",
      "HH_jit_L0.1_variance          0.0\n",
      "HH_jit_L0.1_weight            0.0\n",
      "HH_jit_L1_mean                0.0\n",
      "HpHp_L5_pcc                   0.0\n",
      "HH_jit_L1_weight              0.0\n",
      "HH_jit_L3_variance            0.0\n",
      "HH_jit_L3_mean                0.0\n",
      "HH_jit_L3_weight              0.0\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[14849     0     0]\n",
      " [    0 14860     0]\n",
      " [    0     0 14885]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     14849\n",
      "      gafgyt       1.00      1.00      1.00     14860\n",
      "       mirai       1.00      1.00      1.00     14885\n",
      "\n",
      "    accuracy                           1.00     44594\n",
      "   macro avg       1.00      1.00      1.00     44594\n",
      "weighted avg       1.00      1.00      1.00     44594\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(39339, 115), (39339,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.501\n",
      "HH_jit_L3_mean              0.499\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[3913    0    0]\n",
      " [   0 3891    0]\n",
      " [   0    0 3998]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      3913\n",
      "      gafgyt       1.00      1.00      1.00      3891\n",
      "       mirai       1.00      1.00      1.00      3998\n",
      "\n",
      "    accuracy                           1.00     11802\n",
      "   macro avg       1.00      1.00      1.00     11802\n",
      "weighted avg       1.00      1.00      1.00     11802\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(78200, 115), (78200,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.997\n",
      "HH_jit_L3_mean              0.003\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[11788     0]\n",
      " [    0 11672]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     11788\n",
      "      gafgyt       1.00      1.00      1.00     11672\n",
      "\n",
      "    accuracy                           1.00     23460\n",
      "   macro avg       1.00      1.00      1.00     23460\n",
      "weighted avg       1.00      1.00      1.00     23460\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(525720, 115), (525720,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.500\n",
      "HH_jit_L0.1_mean            0.499\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[52311     0     0]\n",
      " [    0 52769     0]\n",
      " [    0     0 52636]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     52311\n",
      "      gafgyt       1.00      1.00      1.00     52769\n",
      "       mirai       1.00      1.00      1.00     52636\n",
      "\n",
      "    accuracy                           1.00    157716\n",
      "   macro avg       1.00      1.00      1.00    157716\n",
      "weighted avg       1.00      1.00      1.00    157716\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(186462, 115), (186462,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean                0.5\n",
      "HH_jit_L0.1_mean              0.5\n",
      "MI_dir_L5_weight              0.0\n",
      "HH_jit_L1_variance            0.0\n",
      "HpHp_L5_radius                0.0\n",
      "HpHp_L5_magnitude             0.0\n",
      "HpHp_L5_std                   0.0\n",
      "HpHp_L5_mean                  0.0\n",
      "HpHp_L5_weight                0.0\n",
      "HH_jit_L0.01_variance         0.0\n",
      "HH_jit_L0.01_mean             0.0\n",
      "HH_jit_L0.01_weight           0.0\n",
      "HH_jit_L0.1_variance          0.0\n",
      "HH_jit_L0.1_weight            0.0\n",
      "HH_jit_L1_mean                0.0\n",
      "HpHp_L5_pcc                   0.0\n",
      "HH_jit_L1_weight              0.0\n",
      "HH_jit_L3_variance            0.0\n",
      "HH_jit_L3_mean                0.0\n",
      "HH_jit_L3_weight              0.0\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[18587     0     0]\n",
      " [    0 18646     0]\n",
      " [    0     0 18706]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     18587\n",
      "      gafgyt       1.00      1.00      1.00     18646\n",
      "       mirai       1.00      1.00      1.00     18706\n",
      "\n",
      "    accuracy                           1.00     55939\n",
      "   macro avg       1.00      1.00      1.00     55939\n",
      "weighted avg       1.00      1.00      1.00     55939\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(295542, 115), (295542,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.500\n",
      "HH_jit_L1_mean              0.499\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[29478     0     0]\n",
      " [    0 29701     0]\n",
      " [    0     0 29484]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     29478\n",
      "      gafgyt       1.00      1.00      1.00     29701\n",
      "       mirai       1.00      1.00      1.00     29484\n",
      "\n",
      "    accuracy                           1.00     88663\n",
      "   macro avg       1.00      1.00      1.00     88663\n",
      "weighted avg       1.00      1.00      1.00     88663\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(104300, 115), (104300,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.999\n",
      "HH_jit_L3_mean              0.001\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[15642     0]\n",
      " [    0 15648]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     15642\n",
      "      gafgyt       1.00      1.00      1.00     15648\n",
      "\n",
      "    accuracy                           1.00     31290\n",
      "   macro avg       1.00      1.00      1.00     31290\n",
      "weighted avg       1.00      1.00      1.00     31290\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(139755, 115), (139755,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean                0.5\n",
      "HH_jit_L5_mean                0.5\n",
      "HH_jit_L1_variance            0.0\n",
      "HpHp_L5_radius                0.0\n",
      "HpHp_L5_magnitude             0.0\n",
      "HpHp_L5_std                   0.0\n",
      "HpHp_L5_mean                  0.0\n",
      "HpHp_L5_weight                0.0\n",
      "HH_jit_L0.01_variance         0.0\n",
      "HH_jit_L0.01_mean             0.0\n",
      "HH_jit_L0.01_weight           0.0\n",
      "HH_jit_L0.1_variance          0.0\n",
      "HH_jit_L0.1_mean              0.0\n",
      "HH_jit_L0.1_weight            0.0\n",
      "MI_dir_L5_weight              0.0\n",
      "HpHp_L5_pcc                   0.0\n",
      "HH_jit_L1_weight              0.0\n",
      "HH_jit_L3_variance            0.0\n",
      "HH_jit_L3_mean                0.0\n",
      "HH_jit_L3_weight              0.0\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[13999     0     0]\n",
      " [    0 13971     0]\n",
      " [    0     0 13957]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     13999\n",
      "      gafgyt       1.00      1.00      1.00     13971\n",
      "       mirai       1.00      1.00      1.00     13957\n",
      "\n",
      "    accuracy                           1.00     41927\n",
      "   macro avg       1.00      1.00      1.00     41927\n",
      "weighted avg       1.00      1.00      1.00     41927\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(58584, 115), (58584,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.502\n",
      "HH_jit_L0.1_mean            0.498\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "HH_jit_L3_mean              0.000\n",
      "HH_jit_L3_weight            0.000\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[5743    0    0]\n",
      " [   0 5859    0]\n",
      " [   0    0 5974]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      5743\n",
      "      gafgyt       1.00      1.00      1.00      5859\n",
      "       mirai       1.00      1.00      1.00      5974\n",
      "\n",
      "    accuracy                           1.00     17576\n",
      "   macro avg       1.00      1.00      1.00     17576\n",
      "weighted avg       1.00      1.00      1.00     17576\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# with scaling and unbalanced data\n",
    "for k in dataframe:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier_balanced(dataframe[k], k, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(319748, 115), (319748,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.334\n",
      "H_L0.01_variance            0.167\n",
      "MI_dir_L0.01_mean           0.167\n",
      "HH_jit_L1_mean              0.167\n",
      "H_L0.01_mean                0.166\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.635\n",
      "Random Forest: f1-score=0.574\n",
      "Random Forest: Confusion Matrix [[8724    0    0    0    0    2    0    0    0    0    0]\n",
      " [   0    0 8748    0    0    0    0    0    0    0    0]\n",
      " [   2    0 8721    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8746    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 8745    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8675    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8685    0    0    0    0]\n",
      " [   0    0 8740    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8771    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    4    0    0    0    0 8691    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8671]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8726\n",
      "  gafgyt_combo       0.00      0.00      0.00      8748\n",
      "   gafgyt_junk       0.20      1.00      0.33      8723\n",
      "   gafgyt_scan       0.00      0.00      0.00      8746\n",
      "    gafgyt_tcp       1.00      1.00      1.00      8745\n",
      "    gafgyt_udp       1.00      1.00      1.00      8675\n",
      "     mirai_ack       1.00      1.00      1.00      8685\n",
      "    mirai_scan       0.00      0.00      0.00      8740\n",
      "     mirai_syn       0.00      0.00      0.00      8771\n",
      "     mirai_udp       1.00      1.00      1.00      8695\n",
      "mirai_udpplain       1.00      1.00      1.00      8671\n",
      "\n",
      "      accuracy                           0.63     95925\n",
      "     macro avg       0.56      0.64      0.58     95925\n",
      "  weighted avg       0.56      0.63      0.57     95925\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(144243, 115), (144243,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.250\n",
      "MI_dir_L0.01_mean           0.126\n",
      "HH_jit_L1_mean              0.125\n",
      "HH_jit_L3_mean              0.125\n",
      "H_L0.01_mean                0.125\n",
      "MI_dir_L0.01_weight         0.125\n",
      "HH_jit_L5_mean              0.125\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "Random Forest: Accuracy=0.816\n",
      "Random Forest: f1-score=0.770\n",
      "Random Forest: Confusion Matrix [[3931    0    0    0    1    0    0    0    0    0    0]\n",
      " [   0 3954    0    0    0    0    0    0    0    1    0]\n",
      " [   0    0    0    0 4009    0    0    1    0    0    0]\n",
      " [   0    0    0 3834    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 3926    0    0    1    0    1    0]\n",
      " [   0    0    0    0 3964    0    0    1    0    1    0]\n",
      " [   0    0    0    0    0    0 3905    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 3946    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 3906    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 3967    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 3924]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      3932\n",
      "  gafgyt_combo       1.00      1.00      1.00      3955\n",
      "   gafgyt_junk       0.00      0.00      0.00      4010\n",
      "   gafgyt_scan       1.00      1.00      1.00      3834\n",
      "    gafgyt_tcp       0.33      1.00      0.50      3928\n",
      "    gafgyt_udp       0.00      0.00      0.00      3966\n",
      "     mirai_ack       1.00      1.00      1.00      3905\n",
      "    mirai_scan       1.00      1.00      1.00      3946\n",
      "     mirai_syn       1.00      1.00      1.00      3906\n",
      "     mirai_udp       1.00      1.00      1.00      3967\n",
      "mirai_udpplain       1.00      1.00      1.00      3924\n",
      "\n",
      "      accuracy                           0.82     43273\n",
      "     macro avg       0.76      0.82      0.77     43273\n",
      "  weighted avg       0.75      0.82      0.77     43273\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(168720, 115), (168720,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.361\n",
      "H_L0.1_weight               0.208\n",
      "H_L0.01_mean                0.207\n",
      "MI_dir_L0.1_weight          0.207\n",
      "HH_L0.01_radius             0.017\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "Random Forest: Accuracy=0.983\n",
      "Random Forest: f1-score=0.983\n",
      "Random Forest: Confusion Matrix [[8433    0    0    0    0    0]\n",
      " [   1 8415    0    0    0    0]\n",
      " [   0    0 8409    0    0    0]\n",
      " [   2    0    0 8440    0    0]\n",
      " [   1    0    0    0 7537  849]\n",
      " [   1    0    0    0    0 8528]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8433\n",
      "gafgyt_combo       1.00      1.00      1.00      8416\n",
      " gafgyt_junk       1.00      1.00      1.00      8409\n",
      " gafgyt_scan       1.00      1.00      1.00      8442\n",
      "  gafgyt_tcp       1.00      0.90      0.95      8387\n",
      "  gafgyt_udp       0.91      1.00      0.95      8529\n",
      "\n",
      "    accuracy                           0.98     50616\n",
      "   macro avg       0.98      0.98      0.98     50616\n",
      "weighted avg       0.98      0.98      0.98     50616\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(306449, 115), (306449,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L1_mean              0.223\n",
      "HH_jit_L0.1_mean            0.222\n",
      "HH_L0.01_weight             0.111\n",
      "H_L5_mean                   0.111\n",
      "H_L0.1_weight               0.111\n",
      "HH_jit_L5_mean              0.111\n",
      "HH_jit_L0.1_variance        0.111\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "Random Forest: Accuracy=0.908\n",
      "Random Forest: f1-score=0.878\n",
      "Random Forest: Confusion Matrix [[8339    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8407    0    0    0    2    0    0    0    0    0]\n",
      " [   0    0 8320    3    0    2    0    0    0    0    0]\n",
      " [   3    0    0 8386    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8418    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8324    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8440    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 8455    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8309    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8279    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8248]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8339\n",
      "  gafgyt_combo       1.00      1.00      1.00      8409\n",
      "   gafgyt_junk       1.00      1.00      1.00      8325\n",
      "   gafgyt_scan       1.00      1.00      1.00      8389\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8418\n",
      "    gafgyt_udp       0.50      1.00      0.66      8324\n",
      "     mirai_ack       1.00      1.00      1.00      8440\n",
      "    mirai_scan       1.00      1.00      1.00      8455\n",
      "     mirai_syn       1.00      1.00      1.00      8309\n",
      "     mirai_udp       1.00      1.00      1.00      8279\n",
      "mirai_udpplain       1.00      1.00      1.00      8248\n",
      "\n",
      "      accuracy                           0.91     91935\n",
      "     macro avg       0.86      0.91      0.88     91935\n",
      "  weighted avg       0.86      0.91      0.88     91935\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(322267, 115), (322267,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.231\n",
      "MI_dir_L0.1_weight          0.231\n",
      "H_L0.01_variance            0.116\n",
      "HH_jit_L0.1_variance        0.115\n",
      "MI_dir_L0.01_weight         0.115\n",
      "H_L0.1_mean                 0.115\n",
      "HH_jit_L0.01_mean           0.076\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "Random Forest: Accuracy=0.890\n",
      "Random Forest: f1-score=0.859\n",
      "Random Forest: Confusion Matrix [[8796    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8807    1    0    0    0    0    0    0    0    0]\n",
      " [   2    1 8788    0    0    0    0    0    0    0    0]\n",
      " [   1    0    0 8762    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 7038 1807    0    0    0    0    0]\n",
      " [   1    0    0    0    0 8729    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8901    0    1    0    0]\n",
      " [   2    0    1    0    0    0    0 8768    0    0    0]\n",
      " [   4 8822    0    0    0    0    0    0    1    0    0]\n",
      " [   0    0    0    0    0    0    0    0    2 8673    0]\n",
      " [   0    0    0    0    0    0    1    0    0    0 8772]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8796\n",
      "  gafgyt_combo       0.50      1.00      0.67      8808\n",
      "   gafgyt_junk       1.00      1.00      1.00      8791\n",
      "   gafgyt_scan       1.00      1.00      1.00      8763\n",
      "    gafgyt_tcp       1.00      0.80      0.89      8845\n",
      "    gafgyt_udp       0.83      1.00      0.91      8730\n",
      "     mirai_ack       1.00      1.00      1.00      8902\n",
      "    mirai_scan       1.00      1.00      1.00      8771\n",
      "     mirai_syn       0.25      0.00      0.00      8827\n",
      "     mirai_udp       1.00      1.00      1.00      8675\n",
      "mirai_udpplain       1.00      1.00      1.00      8773\n",
      "\n",
      "      accuracy                           0.89     96681\n",
      "     macro avg       0.87      0.89      0.86     96681\n",
      "  weighted avg       0.87      0.89      0.86     96681\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(312367, 115), (312367,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.222\n",
      "MI_dir_L0.01_mean           0.222\n",
      "HH_jit_L0.1_mean            0.111\n",
      "H_L0.01_mean                0.111\n",
      "MI_dir_L0.01_variance       0.111\n",
      "MI_dir_L1_mean              0.111\n",
      "MI_dir_L0.1_mean            0.111\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_weight              0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "Random Forest: Accuracy=0.908\n",
      "Random Forest: f1-score=0.877\n",
      "Random Forest: Confusion Matrix [[8518    0    0    0    0    1    0    1    0    0    0]\n",
      " [   0 8578    6    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8542    0    0    0    0    0    0    0    0]\n",
      " [   2    0    0 8544    0    0    0    0    0    0    0]\n",
      " [   1    0    0    0    0 8619    0    0    0    0    0]\n",
      " [   1    0    0    0    0 8540    0    0    0    0    0]\n",
      " [   0    0    0    0    0    3 8468    0    0    0    0]\n",
      " [   1    0    0    0    0    0    0 8551    0    0    0]\n",
      " [   0    0    0    0    0    1    0    0 8446    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8435    0]\n",
      " [   1    0    0    0    0    0    0    0    0    0 8452]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8520\n",
      "  gafgyt_combo       1.00      1.00      1.00      8584\n",
      "   gafgyt_junk       1.00      1.00      1.00      8542\n",
      "   gafgyt_scan       1.00      1.00      1.00      8546\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8620\n",
      "    gafgyt_udp       0.50      1.00      0.66      8541\n",
      "     mirai_ack       1.00      1.00      1.00      8471\n",
      "    mirai_scan       1.00      1.00      1.00      8552\n",
      "     mirai_syn       1.00      1.00      1.00      8447\n",
      "     mirai_udp       1.00      1.00      1.00      8435\n",
      "mirai_udpplain       1.00      1.00      1.00      8453\n",
      "\n",
      "      accuracy                           0.91     93711\n",
      "     macro avg       0.86      0.91      0.88     93711\n",
      "  weighted avg       0.86      0.91      0.88     93711\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(166188, 115), (166188,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.398\n",
      "H_L0.1_mean                 0.201\n",
      "H_L0.1_weight               0.201\n",
      "MI_dir_L1_weight            0.200\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.998\n",
      "Random Forest: f1-score=0.998\n",
      "Random Forest: Confusion Matrix [[8294    0    0    0    0    0]\n",
      " [   0 8281    1    0    2    0]\n",
      " [   0    0 8261    0    2    0]\n",
      " [   0    0    0 8417    0    0]\n",
      " [   0    0    0    0 8203   94]\n",
      " [   0    0    0    0    0 8302]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8294\n",
      "gafgyt_combo       1.00      1.00      1.00      8284\n",
      " gafgyt_junk       1.00      1.00      1.00      8263\n",
      " gafgyt_scan       1.00      1.00      1.00      8417\n",
      "  gafgyt_tcp       1.00      0.99      0.99      8297\n",
      "  gafgyt_udp       0.99      1.00      0.99      8302\n",
      "\n",
      "    accuracy                           1.00     49857\n",
      "   macro avg       1.00      1.00      1.00     49857\n",
      "weighted avg       1.00      1.00      1.00     49857\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(306075, 115), (306075,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_variance            0.143\n",
      "HH_jit_L3_mean              0.143\n",
      "H_L0.01_weight              0.143\n",
      "HH_jit_L0.1_mean            0.143\n",
      "HH_jit_L1_mean              0.143\n",
      "HH_jit_L0.01_mean           0.142\n",
      "HH_L1_weight                0.142\n",
      "HH_jit_L5_mean              0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_L0.01_magnitude          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_L0.01_radius             0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L5_variance          0.000\n",
      "Random Forest: Accuracy=0.724\n",
      "Random Forest: f1-score=0.648\n",
      "Random Forest: Confusion Matrix [[8336    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8400    0    0    0    3    0    0    0    0    0]\n",
      " [   4    0 8305    0    0    2    0    0    0    0    0]\n",
      " [8395    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8465    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8315    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8330    0    0    0    0]\n",
      " [   0    0    0    0    0 8476    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8231    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8315    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8246]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       0.50      1.00      0.66      8336\n",
      "  gafgyt_combo       1.00      1.00      1.00      8403\n",
      "   gafgyt_junk       1.00      1.00      1.00      8311\n",
      "   gafgyt_scan       0.00      0.00      0.00      8395\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8465\n",
      "    gafgyt_udp       0.33      1.00      0.50      8315\n",
      "     mirai_ack       1.00      1.00      1.00      8330\n",
      "    mirai_scan       0.00      0.00      0.00      8476\n",
      "     mirai_syn       1.00      1.00      1.00      8231\n",
      "     mirai_udp       1.00      1.00      1.00      8315\n",
      "mirai_udpplain       1.00      1.00      1.00      8246\n",
      "\n",
      "      accuracy                           0.72     91823\n",
      "     macro avg       0.62      0.73      0.65     91823\n",
      "  weighted avg       0.62      0.72      0.65     91823\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(214808, 115), (214808,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.178\n",
      "HH_jit_L5_mean              0.118\n",
      "HH_jit_L0.01_weight         0.118\n",
      "HH_L0.01_weight             0.117\n",
      "H_L0.1_mean                 0.117\n",
      "H_L0.1_weight               0.117\n",
      "HH_jit_L1_mean              0.117\n",
      "HH_jit_L0.1_mean            0.117\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "Random Forest: Accuracy=0.879\n",
      "Random Forest: f1-score=0.847\n",
      "Random Forest: Confusion Matrix [[5904    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 5817    0    0    0    0    0    1    0    0    0]\n",
      " [   0    0 5846    0    0    0    0    2    0    2    0]\n",
      " [   1    0    0 5758    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 3978 1852    0    0    0    0    0]\n",
      " [   0    0    0    0    2 5840    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 5953    0]\n",
      " [   0    0    0    0    0    0    0 5875    0    2    0]\n",
      " [   0    0    0    0    0    0    0    0 5782    0    0]\n",
      " [   0    0    0    0    0    0    0    2    0 5916    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 5910]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      5904\n",
      "  gafgyt_combo       1.00      1.00      1.00      5818\n",
      "   gafgyt_junk       1.00      1.00      1.00      5850\n",
      "   gafgyt_scan       1.00      1.00      1.00      5759\n",
      "    gafgyt_tcp       1.00      0.68      0.81      5830\n",
      "    gafgyt_udp       0.76      1.00      0.86      5842\n",
      "     mirai_ack       0.00      0.00      0.00      5953\n",
      "    mirai_scan       1.00      1.00      1.00      5877\n",
      "     mirai_syn       1.00      1.00      1.00      5782\n",
      "     mirai_udp       0.50      1.00      0.67      5918\n",
      "mirai_udpplain       1.00      1.00      1.00      5910\n",
      "\n",
      "      accuracy                           0.88     64443\n",
      "     macro avg       0.84      0.88      0.85     64443\n",
      "  weighted avg       0.84      0.88      0.85     64443\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# without scaling and unbalanced data multi label\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier_balanced(dataframe_multi[k], k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------xxxxxxx----------------------\n",
      "DanminDoorBell_multi \n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1018298, 115), (1018298,)\n",
      "Imblearn Shape:(319748, 115), (319748,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.01_variance              0.2\n",
      "H_L0.01_mean                  0.2\n",
      "HH_jit_L5_mean                0.2\n",
      "MI_dir_L0.01_mean             0.2\n",
      "HH_jit_L1_mean                0.2\n",
      "HH_jit_L0.1_mean              0.0\n",
      "HpHp_L5_covariance            0.0\n",
      "HpHp_L5_radius                0.0\n",
      "HpHp_L5_magnitude             0.0\n",
      "HpHp_L5_std                   0.0\n",
      "HpHp_L5_mean                  0.0\n",
      "HpHp_L5_weight                0.0\n",
      "HH_jit_L0.01_variance         0.0\n",
      "HH_jit_L0.01_mean             0.0\n",
      "HH_jit_L0.01_weight           0.0\n",
      "HH_jit_L0.1_variance          0.0\n",
      "HH_jit_L1_variance            0.0\n",
      "HH_jit_L0.1_weight            0.0\n",
      "HpHp_L3_weight                0.0\n",
      "HH_jit_L1_weight              0.0\n",
      "Random Forest: Accuracy=0.544\n",
      "Random Forest: f1-score=0.453\n",
      "Random Forest: Confusion Matrix [[8724    0    0    0    0    2    0    0    0    0    0]\n",
      " [   0    0 8748    0    0    0    0    0    0    0    0]\n",
      " [   2    0 8721    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8746    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8745    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8675    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8685    0    0    0    0]\n",
      " [   0    0 8740    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8771    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8695    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8671]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8726\n",
      "  gafgyt_combo       0.00      0.00      0.00      8748\n",
      "   gafgyt_junk       0.20      1.00      0.33      8723\n",
      "   gafgyt_scan       0.00      0.00      0.00      8746\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8745\n",
      "    gafgyt_udp       0.50      1.00      0.66      8675\n",
      "     mirai_ack       1.00      1.00      1.00      8685\n",
      "    mirai_scan       0.00      0.00      0.00      8740\n",
      "     mirai_syn       0.00      0.00      0.00      8771\n",
      "     mirai_udp       1.00      1.00      1.00      8695\n",
      "mirai_udpplain       1.00      1.00      1.00      8671\n",
      "\n",
      "      accuracy                           0.54     95925\n",
      "     macro avg       0.43      0.55      0.45     95925\n",
      "  weighted avg       0.43      0.54      0.45     95925\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ecobee_Thermostat_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(835876, 115), (835876,)\n",
      "Imblearn Shape:(144243, 115), (144243,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L5_mean              0.375\n",
      "MI_dir_L0.01_mean           0.251\n",
      "HH_jit_L0.1_mean            0.125\n",
      "H_L0.01_mean                0.125\n",
      "HH_jit_L1_mean              0.125\n",
      "MI_dir_L5_weight            0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.816\n",
      "Random Forest: f1-score=0.770\n",
      "Random Forest: Confusion Matrix [[3931    1    0    0    0    0    0    0    0    0    0]\n",
      " [   0 3954    0    0    0    0    0    0    0    1    0]\n",
      " [   0 4008    0    0    1    0    0    1    0    0    0]\n",
      " [   0    0    0 3834    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 3926    0    0    1    0    1    0]\n",
      " [   0 3964    0    0    0    0    0    1    0    1    0]\n",
      " [   0    0    0    0    0    0 3905    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0 3946    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 3906    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 3967    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 3924]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      3932\n",
      "  gafgyt_combo       0.33      1.00      0.50      3955\n",
      "   gafgyt_junk       0.00      0.00      0.00      4010\n",
      "   gafgyt_scan       1.00      1.00      1.00      3834\n",
      "    gafgyt_tcp       1.00      1.00      1.00      3928\n",
      "    gafgyt_udp       0.00      0.00      0.00      3966\n",
      "     mirai_ack       1.00      1.00      1.00      3905\n",
      "    mirai_scan       1.00      1.00      1.00      3946\n",
      "     mirai_syn       1.00      1.00      1.00      3906\n",
      "     mirai_udp       1.00      1.00      1.00      3967\n",
      "mirai_udpplain       1.00      1.00      1.00      3924\n",
      "\n",
      "      accuracy                           0.82     43273\n",
      "     macro avg       0.76      0.82      0.77     43273\n",
      "  weighted avg       0.75      0.82      0.77     43273\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Ennio_Doorbell_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(355500, 115), (355500,)\n",
      "Imblearn Shape:(168720, 115), (168720,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_weight               0.523\n",
      "H_L0.01_mean                0.250\n",
      "HH_jit_L0.01_weight         0.225\n",
      "H_L0.1_mean                 0.002\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "Random Forest: Accuracy=0.831\n",
      "Random Forest: f1-score=0.776\n",
      "Random Forest: Confusion Matrix [[8433    0    0    0    0    0]\n",
      " [   1 8415    0    0    0    0]\n",
      " [   0    0 8409    0    0    0]\n",
      " [   2    0    0 8440    0    0]\n",
      " [   1    0    0    0 8386    0]\n",
      " [   1    0    0    0 8528    0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8433\n",
      "gafgyt_combo       1.00      1.00      1.00      8416\n",
      " gafgyt_junk       1.00      1.00      1.00      8409\n",
      " gafgyt_scan       1.00      1.00      1.00      8442\n",
      "  gafgyt_tcp       0.50      1.00      0.66      8387\n",
      "  gafgyt_udp       0.00      0.00      0.00      8529\n",
      "\n",
      "    accuracy                           0.83     50616\n",
      "   macro avg       0.75      0.83      0.78     50616\n",
      "weighted avg       0.75      0.83      0.78     50616\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Philips_B120N10_Baby_Monitor_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(1098677, 115), (1098677,)\n",
      "Imblearn Shape:(306449, 115), (306449,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.1_mean            0.222\n",
      "HH_jit_L5_mean              0.112\n",
      "HH_L0.01_weight             0.111\n",
      "MI_dir_L1_mean              0.111\n",
      "H_L0.1_weight               0.111\n",
      "HH_jit_L0.01_mean           0.111\n",
      "HH_jit_L0.1_variance        0.111\n",
      "HH_jit_L1_mean              0.111\n",
      "HH_jit_L5_weight            0.000\n",
      "HH_L0.01_pcc                0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_L0.01_std                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_L0.01_magnitude          0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_L0.01_radius             0.000\n",
      "Random Forest: Accuracy=0.908\n",
      "Random Forest: f1-score=0.877\n",
      "Random Forest: Confusion Matrix [[8339    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8407    0    0    2    0    0    0    0    0    0]\n",
      " [   0    0 8320    3    2    0    0    0    0    0    0]\n",
      " [   3    0    0 8386    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 8418    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8324    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8440    0    0    0    0]\n",
      " [   0    0    0    0 8455    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8309    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8279    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8248]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8339\n",
      "  gafgyt_combo       1.00      1.00      1.00      8409\n",
      "   gafgyt_junk       1.00      1.00      1.00      8325\n",
      "   gafgyt_scan       1.00      1.00      1.00      8389\n",
      "    gafgyt_tcp       0.50      1.00      0.67      8418\n",
      "    gafgyt_udp       1.00      1.00      1.00      8324\n",
      "     mirai_ack       1.00      1.00      1.00      8440\n",
      "    mirai_scan       0.00      0.00      0.00      8455\n",
      "     mirai_syn       1.00      1.00      1.00      8309\n",
      "     mirai_udp       1.00      1.00      1.00      8279\n",
      "mirai_udpplain       1.00      1.00      1.00      8248\n",
      "\n",
      "      accuracy                           0.91     91935\n",
      "     macro avg       0.86      0.91      0.88     91935\n",
      "  weighted avg       0.86      0.91      0.88     91935\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_737E_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(828260, 115), (828260,)\n",
      "Imblearn Shape:(322267, 115), (322267,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_weight         0.250\n",
      "H_L0.01_mean                0.250\n",
      "H_L0.1_weight               0.125\n",
      "MI_dir_L0.01_variance       0.125\n",
      "MI_dir_L0.01_mean           0.125\n",
      "HH_jit_L0.1_variance        0.125\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "Random Forest: Accuracy=0.817\n",
      "Random Forest: f1-score=0.756\n",
      "Random Forest: Confusion Matrix [[8796    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8807    1    0    0    0    0    0    0    0    0]\n",
      " [   2    1 8788    0    0    0    0    0    0    0    0]\n",
      " [   1    0    0 8762    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8845    0    0    0    0    0]\n",
      " [   1    0    0    0    0 8729    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8901    0    1    0    0]\n",
      " [   2    0    1    0    0    0    0 8768    0    0    0]\n",
      " [   4 8822    0    0    0    0    0    0    1    0    0]\n",
      " [   0    0    0    0    0    0    0    0    2 8673    0]\n",
      " [   0    0    0    0    0    0    1    0    0    0 8772]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8796\n",
      "  gafgyt_combo       0.50      1.00      0.67      8808\n",
      "   gafgyt_junk       1.00      1.00      1.00      8791\n",
      "   gafgyt_scan       1.00      1.00      1.00      8763\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8845\n",
      "    gafgyt_udp       0.50      1.00      0.66      8730\n",
      "     mirai_ack       1.00      1.00      1.00      8902\n",
      "    mirai_scan       1.00      1.00      1.00      8771\n",
      "     mirai_syn       0.25      0.00      0.00      8827\n",
      "     mirai_udp       1.00      1.00      1.00      8675\n",
      "mirai_udpplain       1.00      1.00      1.00      8773\n",
      "\n",
      "      accuracy                           0.82     96681\n",
      "     macro avg       0.75      0.82      0.76     96681\n",
      "  weighted avg       0.75      0.82      0.76     96681\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Provision_PT_838_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(836891, 115), (836891,)\n",
      "Imblearn Shape:(312367, 115), (312367,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.222\n",
      "H_L0.1_weight               0.222\n",
      "HH_jit_L0.1_mean            0.111\n",
      "H_L0.01_variance            0.111\n",
      "H_L0.01_mean                0.111\n",
      "MI_dir_L1_mean              0.111\n",
      "MI_dir_L0.01_weight         0.111\n",
      "HpHp_L5_mean                0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HpHp_L5_weight              0.000\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "Random Forest: Accuracy=0.908\n",
      "Random Forest: f1-score=0.877\n",
      "Random Forest: Confusion Matrix [[8518    0    0    0    0    2    0    0    0    0    0]\n",
      " [   0 8579    5    0    0    0    0    0    0    0    0]\n",
      " [   0    0 8542    0    0    0    0    0    0    0    0]\n",
      " [   2    0    0 8544    0    0    0    0    0    0    0]\n",
      " [   1    0    0    0    0    0    0 8619    0    0    0]\n",
      " [   1    0    0    0    0 8540    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8468    3    0    0    0]\n",
      " [   1    0    0    0    0    0    0 8551    0    0    0]\n",
      " [   0    0    0    0    0    0    0    1 8446    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8435    0]\n",
      " [   1    0    0    0    0    0    0    0    0    0 8452]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      8520\n",
      "  gafgyt_combo       1.00      1.00      1.00      8584\n",
      "   gafgyt_junk       1.00      1.00      1.00      8542\n",
      "   gafgyt_scan       1.00      1.00      1.00      8546\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8620\n",
      "    gafgyt_udp       1.00      1.00      1.00      8541\n",
      "     mirai_ack       1.00      1.00      1.00      8471\n",
      "    mirai_scan       0.50      1.00      0.66      8552\n",
      "     mirai_syn       1.00      1.00      1.00      8447\n",
      "     mirai_udp       1.00      1.00      1.00      8435\n",
      "mirai_udpplain       1.00      1.00      1.00      8453\n",
      "\n",
      "      accuracy                           0.91     93711\n",
      "     macro avg       0.86      0.91      0.88     93711\n",
      "  weighted avg       0.86      0.91      0.88     93711\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "Samsung_SNH_1011_N_Webcam_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(375222, 115), (375222,)\n",
      "Imblearn Shape:(166188, 115), (166188,)\n",
      "                       importance\n",
      "feature                          \n",
      "H_L0.1_mean                 0.250\n",
      "H_L0.1_weight               0.250\n",
      "MI_dir_L0.01_weight         0.250\n",
      "MI_dir_L1_weight            0.249\n",
      "MI_dir_L5_weight            0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.01_mean           0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L3_weight              0.000\n",
      "HH_jit_L1_mean              0.000\n",
      "Random Forest: Accuracy=0.833\n",
      "Random Forest: f1-score=0.778\n",
      "Random Forest: Confusion Matrix [[8294    0    0    0    0    0]\n",
      " [   0 8281    1    0    2    0]\n",
      " [   0    0 8261    0    2    0]\n",
      " [   1    0    0 8416    0    0]\n",
      " [   0    0    0    0 8297    0]\n",
      " [   0    0    0    0 8302    0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      8294\n",
      "gafgyt_combo       1.00      1.00      1.00      8284\n",
      " gafgyt_junk       1.00      1.00      1.00      8263\n",
      " gafgyt_scan       1.00      1.00      1.00      8417\n",
      "  gafgyt_tcp       0.50      1.00      0.67      8297\n",
      "  gafgyt_udp       0.00      0.00      0.00      8302\n",
      "\n",
      "    accuracy                           0.83     49857\n",
      "   macro avg       0.75      0.83      0.78     49857\n",
      "weighted avg       0.75      0.83      0.78     49857\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1002_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(863056, 115), (863056,)\n",
      "Imblearn Shape:(306075, 115), (306075,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L0.01_mean           0.286\n",
      "H_L0.01_variance            0.143\n",
      "HH_jit_L3_mean              0.143\n",
      "H_L0.01_weight              0.143\n",
      "HH_jit_L1_mean              0.143\n",
      "HH_L1_weight                0.142\n",
      "HH_jit_L5_weight            0.000\n",
      "HH_jit_L5_mean              0.000\n",
      "HpHp_L5_covariance          0.000\n",
      "HH_L0.01_magnitude          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_L0.01_radius             0.000\n",
      "HH_jit_L0.01_weight         0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L0.1_mean            0.000\n",
      "Random Forest: Accuracy=0.724\n",
      "Random Forest: f1-score=0.648\n",
      "Random Forest: Confusion Matrix [[8336    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 8400    0    0    0    3    0    0    0    0    0]\n",
      " [   4    0 8305    0    0    2    0    0    0    0    0]\n",
      " [8395    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8465    0    0    0    0    0]\n",
      " [   0    0    0    0    0 8315    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0 8330    0    0    0    0]\n",
      " [   0    0    0    0    0 8476    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0 8231    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 8315    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 8246]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       0.50      1.00      0.66      8336\n",
      "  gafgyt_combo       1.00      1.00      1.00      8403\n",
      "   gafgyt_junk       1.00      1.00      1.00      8311\n",
      "   gafgyt_scan       0.00      0.00      0.00      8395\n",
      "    gafgyt_tcp       0.00      0.00      0.00      8465\n",
      "    gafgyt_udp       0.33      1.00      0.50      8315\n",
      "     mirai_ack       1.00      1.00      1.00      8330\n",
      "    mirai_scan       0.00      0.00      0.00      8476\n",
      "     mirai_syn       1.00      1.00      1.00      8231\n",
      "     mirai_udp       1.00      1.00      1.00      8315\n",
      "mirai_udpplain       1.00      1.00      1.00      8246\n",
      "\n",
      "      accuracy                           0.72     91823\n",
      "     macro avg       0.62      0.73      0.65     91823\n",
      "  weighted avg       0.62      0.72      0.65     91823\n",
      "\n",
      "---------------------- Done ---------------------\n",
      "----------------------xxxxxxx----------------------\n",
      "SimpleHome_XCS7_1003_WHT_Security_Camera_multi\n",
      "----------------------xxxxxxx----------------------\n",
      "Original Shape:(850826, 115), (850826,)\n",
      "Imblearn Shape:(214808, 115), (214808,)\n",
      "                       importance\n",
      "feature                          \n",
      "HH_jit_L3_mean              0.222\n",
      "HH_jit_L0.01_mean           0.112\n",
      "HH_L0.01_weight             0.111\n",
      "H_L0.1_mean                 0.111\n",
      "H_L0.1_weight               0.111\n",
      "HH_jit_L0.01_weight         0.111\n",
      "HH_jit_L1_mean              0.111\n",
      "HH_jit_L0.1_mean            0.111\n",
      "HH_jit_L0.1_weight          0.000\n",
      "HpHp_L5_radius              0.000\n",
      "HpHp_L5_magnitude           0.000\n",
      "HpHp_L5_std                 0.000\n",
      "HpHp_L5_mean                0.000\n",
      "HpHp_L5_weight              0.000\n",
      "HH_jit_L0.01_variance       0.000\n",
      "HH_jit_L0.1_variance        0.000\n",
      "HH_jit_L1_variance          0.000\n",
      "HpHp_L5_pcc                 0.000\n",
      "HH_jit_L1_weight            0.000\n",
      "HH_jit_L3_variance          0.000\n",
      "Random Forest: Accuracy=0.907\n",
      "Random Forest: f1-score=0.877\n",
      "Random Forest: Confusion Matrix [[5904    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 5817    0    0    0    0    0    1    0    0    0]\n",
      " [   0    0 5846    0    0    0    0    2    0    2    0]\n",
      " [   1    0    0 5758    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0 5830    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 5842    0    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0 5953    0]\n",
      " [   0    0    0    0    0    0    0 5875    0    2    0]\n",
      " [   0    0    0    0    0    0    0    0 5782    0    0]\n",
      " [   0    0    0    0    0    0    0    2    0 5916    0]\n",
      " [   0    0    0    0    0    0    0    0    0    0 5910]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00      5904\n",
      "  gafgyt_combo       1.00      1.00      1.00      5818\n",
      "   gafgyt_junk       1.00      1.00      1.00      5850\n",
      "   gafgyt_scan       1.00      1.00      1.00      5759\n",
      "    gafgyt_tcp       1.00      1.00      1.00      5830\n",
      "    gafgyt_udp       1.00      1.00      1.00      5842\n",
      "     mirai_ack       0.00      0.00      0.00      5953\n",
      "    mirai_scan       1.00      1.00      1.00      5877\n",
      "     mirai_syn       1.00      1.00      1.00      5782\n",
      "     mirai_udp       0.50      1.00      0.67      5918\n",
      "mirai_udpplain       1.00      1.00      1.00      5910\n",
      "\n",
      "      accuracy                           0.91     64443\n",
      "     macro avg       0.86      0.91      0.88     64443\n",
      "  weighted avg       0.86      0.91      0.88     64443\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "# without scaling and unbalanced data multi label\n",
    "for k in dataframe_multi:\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    print(k)\n",
    "    print(\"----------------------xxxxxxx----------------------\")\n",
    "    dt_classifier_balanced(dataframe_multi[k], k, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Scaling & PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_pca(df, n_components):\n",
    "    X = df.drop(['label','device'], axis=1).values\n",
    "    \n",
    "    X_std = StandardScaler().fit_transform(X)\n",
    "    pca = PCA(n_components=n_components)\n",
    "    X_pca = pca.fit_transform(X_std)\n",
    "    \n",
    "    return X_pca"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_doorbell_df_pca = scaling_pca(danmini_doorbell_df, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "original shape:    (1018298, 117)\n",
      "transformed shape: (1018298, 10)\n"
     ]
    }
   ],
   "source": [
    "print(\"original shape:   \", danmini_doorbell_df.shape)\n",
    "print(\"transformed shape:\", danmini_doorbell_df_pca.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy86wFpkAAAACXBIWXMAAAsTAAALEwEAmpwYAAAoYklEQVR4nO3deZhcVZ3/8fenu5N09qRJCCELCRg2ZTVAcEFQVMAFt0dBwWVUZBQBHZ3BGcfR8TfjiDqMM6ARFUFHYVxZnMimAjODQBIMZAMSQiCdhKSzd3d67+/vj3s7VJpeboeurq6qz+t56qm6t+699T1V3fWte8495ygiMDOz8lVR6ADMzKywnAjMzMqcE4GZWZlzIjAzK3NOBGZmZa6q0AEM1JQpU2LOnDmFDsPMrKgsXbp0W0RM7em5oksEc+bMYcmSJYUOw8ysqEh6trfnXDVkZlbmnAjMzMqcE4GZWZlzIjAzK3NOBGZmZS5viUDSDZK2SlrRy/OS9O+S1kp6XNLJ+YrFzMx6l88zghuBc/p4/lxgXnq7BPhuHmMxM7Ne5K0fQUQ8IGlOH5ucD/w4knGwH5I0SdL0iNicr5isOEUEbR1BS3sHre2dtHZ00treSVtHJ63tQVtHJ+2dnbR3BB2dQVtn0JGz3N4ZdEZ664TOCCJI16XLyQsRQGdnch9Beh9pHJBumfNcGiNB9xHdI/ZfF/s998J+PZe5h3UDe9v65uHni9L8OTWccWSPfcJekkJ2KJsBbMhZrk3XvSgRSLqE5KyB2bNnD0lwdmA6O4P65nZ2NbWyp6mdPc1t7Glqo76lnfrmdhpbkltDSztNrR00trazt7WD5rYOmto6aG7rpLmtg5b2Tlq67ts7C12skiQVOgIbqEtfd0TJJYKe/gx7/JkSEdcD1wPMnz/fP2UKoKMzqKtvYfPuJrbsaeb53c1sqW9hy55mtu5poa6+hbqGFnbtbaWzn09oZFUF40ZVMXZUJWNHVjF6ZCWjR1QyvnoE1SMqqK6qZNSICkZVVTKqqoJRI5L7kZUVjKxKbiO6HleKEZUVVFVWUFWh5FYpKiuS5cp0XUWFqJColJCSL8HKCiFEhQCx77EkRLKNEEorUJN12veHu+/5nL/k/R6nz+X+oStnA/WwD71sa5ZPhUwEtcCsnOWZwKYCxWIkVRlb61t48vl61mxtYO3Wep7e2siGnXvZWt9CR7dv+BGVYuq4URw8oZrDDhrDK+dMpmbMSCaNGcGkMSOZOHoE46urmFCd3I8bVcW46ipGVPpiNbPhpJCJ4HbgMkm3AKcBu90+MHTaOjp58vl6lm/czerNe3hicz1Pbqlnd1Pbvm0mjxnByw4ex+lHHMShE0dzyMRqpk+sZtqEag6ZWE3NmJFUVPhXq1mxy1sikHQzcCYwRVIt8A/ACICIWAgsAs4D1gJ7gY/kKxaD3XvbeGT9DpY8u4NHn93J47W799W9jxtVxZHTxnHecdM5+pDxzJs2jiOnjWfKuFEFjtrMhkI+rxq6sJ/nA/hUvl6/3DW1dvDwM9v53zXbePDp7ax+fg8RSXXOK2ZM5KIFh3HCrEmcMHMis2vGuD7arIwV3TDU1rsNO/Zy7+ot/OGJrTy8bgetHZ2MrKzg5MMmceUbjmTB4TWcMGsS1SMqCx2qmQ0jTgRF7um6BhY9vpnfrXieVZv3AHDE1LF86FWH8dp5UzllTg2jR/qL38x650RQhPa2tvPbxzbzX0s2sPTZnQC88rDJfPEtx3D2MdOYM2VsgSM0s2LiRFBEmlo7+MlD6/ne/evY3tjKEVPH8oVzj+b8E2dwyMTqQodnZkXKiaAIdHQGP1+ygW/d/RTbGlp47bwpfPr18zhlzmQ38prZS+ZEMMwtXr+DL9++kpWb9nDKnMl896KTOWVOTaHDMrMS4kQwTO3e28bXfreaWxZv4NCJ1fzHhSfx1uOn+wzAzAadE8Ew9Lvlm/n721ayc28rnzjjcK44ex5jRvqjMrP88LfLMLJ7bxtfun0Fty3bxCtmTODGj5zCK2ZMLHRYZlbinAiGiRUbd/PRmxazvaGVK8+ex6fOepkHZzOzIeFEMAws27CLi3/4MBOqR/CbT76a42b6LMDMho4TQYEtfXYHH7phMTVjR/Kzj5/GzMljCh2SmZUZJ4IC+vNzO/ngDx/h4AnV/OzjpzF94uhCh2RmZciJoEBWbtrNh254hIPGjeLmjy9wz2AzKxi3RhbAmi31XPzDRxg3qoqfffw0JwEzKygngiG2vaGFD/9oMZUV4mcfX+A2ATMrOFcNDaG2jk4++dNHqWto4RefON2jhJrZsOAzgiH0lTtW8vAzO7j63cdzwqxJhQ7HzAzImAgkHSbp7PTxaEnj8xtW6fnl0lr+86Hn+MTrDucdJ80odDhmZvv0mwgkfRz4JfC9dNVM4NY8xlRy1m9r5Eu3reC0uTX89ZuPLnQ4Zmb7yXJG8Cng1cAegIhYAxycz6BKSVtHJ1fc8mdGVFZwzftOpLLCo4ea2fCSJRG0RERr14KkKiDyF1Jpueaep3isdjdfe9dxHDrJHcbMbPjJkgjul/S3wGhJbwR+AdyR5eCSzpH0pKS1kq7q4fnJkn4j6XFJj0h6xcDCH97Wbm3gew+s473zZ3LecdMLHY6ZWY+yJIKrgDpgOfAJYBHwxf52klQJXAecCxwLXCjp2G6b/S2wLCKOBz4IfDt76MPf1Xc+wegRlVx17jGFDsXMrFdZ+hGMBm6IiO/Dvi/40cDefvY7FVgbEevS/W4BzgdW5WxzLPA1gIh4QtIcSdMiYsvAijH8LFm/g7tXbeFzbzqSmrEjCx2OmVmvspwR/J7ki7/LaODeDPvNADbkLNem63I9BrwLQNKpwGEkVyXtR9IlkpZIWlJXV5fhpQsrIvja757g4PGj+IvXzC10OGZmfcqSCKojoqFrIX2cZVyEni6P6d7I/C/AZEnLgE8DfwbaX7RTxPURMT8i5k+dOjXDSxfWPau2sPTZnVx59pGeYtLMhr0s31KNkk6OiEcBJL0SaMqwXy0wK2d5JrApd4OI2AN8JD2ugGfSW9GKCK65dw2HTxnLe+e/6OTGzGzYyZIIrgR+IanrS3w68L4M+y0G5kmaC2wELgDen7uBpEnA3vTy1I8BD6TJoWg9sGYbqzfv4er3HE+Vp5o0syLQbyKIiMWSjgaOIqnueSIi2jLs1y7pMuAuoJKkwXmlpEvT5xcCxwA/ltRB0oj80QMvyvDw3fvWcsiEat5xooeRMLPikLUC+xRgTrr9SZKIiB/3t1NELCK53DR33cKcx38C5mWOdphbtmEXD63bwRffcgwjq3w2YGbFod9EIOknwBHAMqAjXR1Av4mg3Cy872kmVFdxwamzCx2KmVlmWc4I5gPHRoSHlejDuroG7lr1PJed9TLGjfKVQmZWPLLUX6wADsl3IMXul0trqZC4+PTDCh2KmdmAZPnpOgVYJekRoKVrZUS8PW9RFZnOzuC2ZZt4zcumcPB4zz9sZsUlSyL4cr6DKHZLn9vJxl1NfP7NRxU6FDOzActy+ej9QxFIMfvNnzcyekQlbzx2WqFDMTMbsCwzlC2QtFhSg6RWSR2SirrT12Bqbe9k0fLNvOnl0xjrRmIzK0JZGouvBS4E1pAMOPexdJ0B9z9Vx669be5AZmZFK9NP2IhYK6kyIjqAH0l6MM9xFY1b/7yRmrEjec28KYUOxczsgGRJBHsljQSWSboa2AyMzW9YxaGhpZ17V2/hfafMYoTHFTKzIpXl2+tikrGCLgMaSUYUfXc+gyoWv1+9hZb2Tt52wqGFDsXM7IBluWro2fRhE/CV/IZTXH77+GYOmVDNK2dPLnQoZmYHrNdEIOnnEfFeSct58YQypPMMl6365jbuf7KOixYcRkVFT3PwmJkVh77OCK5I7986FIEUm3tXb6G1o5O3HD+90KGYmb0kvSaCiNicTlT/w4g4ewhjKgr//fhmDp1YzUmzJhU6FDOzl6TPxuL0ctG9kiYOUTxFYXdTGw88tY3zjpvuaiEzK3pZLh9tBpZLuofkqiEAIuLyvEU1zN2zytVCZlY6siSC/05vlrpzxWZmTBrNia4WMrMSkOXy0ZuGIpBi0dTawf+s2caFp85GcrWQmRW/LFNVzgO+BhwL7BtsPyIOz2Ncw9aDT2+jpb2TNxxzcKFDMTMbFFl6Fv8I+C7QDpxFMlfxT/IZ1HB27+qtjB1ZyalzawodipnZoMiSCEZHxO8BRcSzEfFl4PX5DWt4igj+8MQWzjhyKqOqKgsdjpnZoMiSCJolVQBrJF0m6Z1ApnoRSedIelLSWklX9fD8REl3SHpM0kpJHxlg/ENqxcY9bNnTwhuO8QQ0ZlY6siSCK4ExwOXAK4GLgA/1t1PaGe064FyS9oULJR3bbbNPAasi4gTgTOBb6Uinw9K9q7cgwVlHTS10KGZmgybL5aPtEdEANAAD+cV+KrA2ItYBSLoFOB9YlbNNAOOVXH4zDthB0hYxLP3hia2cPHsyB40bVehQzMwGTZYzgn+V9ISkr0p6+QCOPQPYkLNcm67LdS1wDLAJWA5cERGd3Q8k6RJJSyQtqaurG0AIg2fLnmaWb9ztq4XMrOT0mwgi4iySaps64HpJyyV9McOxe7rIvvsopm8GlgGHAicC10qa0EMM10fE/IiYP3VqYapl7n8qSUCvP9qJwMxKS6ZptSLi+Yj4d+BSki/uL2XYrZZkEpsuM0l++ef6CPDrSKwFngGOzhLTUHt43Q5qxo7kqGnjCx2Kmdmg6jcRSDpG0pclrSCpynmQ5Eu9P4uBeZLmpg3AFwC3d9vmOeAN6etMA44C1g0g/iHz8DPbOXVOjXsTm1nJydJY/CPgZuBNEdH9F32vIqJd0mXAXSRTXd4QESslXZo+vxD4KnBjOvmNgL+JiG0DLUS+bdzVRO3OJj76mrmFDsXMbNBlGWtowYEePCIWAYu6rVuY83gT8KYDPf5QeXjddgBOm3tQgSMxMxt8mdoIyt3D63YwcfQIjj7E7QNmVnqcCDJ4+JntnDKnxpPQmFlJciLox5Y9zazfvpfTPMicmZWoXtsIJN3Bi6/73yci3p6XiIaZh7raBw53IjCz0tRXY/E30/t3AYcA/5kuXwisz2NMw8rDz+xg3Kgqjp3+on5uZmYloddEEBH3A0j6akSckfPUHZIeyHtkw8TD67Yzf85kqipdi2ZmpSnLt9tUSftmI5M0FyiL4Tfrm9t4uq6R+YdNLnQoZmZ5k6VD2WeA+yR19fidA3wibxENI2u3NgBwpIeVMLMSlqVD2Z3pvMVdYwA9EREt+Q1reFizJUkE85wIzKyEZRlraAzweeCyiHgMmC3prXmPbBh4aks9o6oqmF0zptChmJnlTdbJ61uB09PlWuD/5S2iYeSprQ0cMXUcle5IZmYlLEsiOCIirgbaACKiiZ7nGig5a7fUc+S0cYUOw8wsr7IkglZJo0k7l0k6Aij5NoL65jY27W52+4CZlbwsVw39A3AnMEvST4FXAx/OZ1DDwZr0iqF5B/uMwMxKW5arhu6R9CiwgKRK6IrhOGfAYFu7xZeOmll5yHJGAFAN7Ey3P1YSEVHSvYu7rhia5SuGzKzE9ZsIJH0deB+wEuhMVwdQ2olgawMvO9hXDJlZ6ctyRvAO4Khy6UTWZc2Weg89bWZlIctVQ+uAEfkOZDipb25js68YMrMykeWMYC+wTNLvyblsNCIuz1tUBbbGYwyZWRnJkghuT29lY82WegB3JjOzspDl8tGbhiKQ4WTNlgZGVVUwc7KvGDKz0tdrG4Gkn6f3yyU93v2W5eCSzpH0pKS1kq7q4fnPS1qW3lZI6pBU8BbaZ7Y1MnfKWF8xZGZloa8zgivS+wMaaVRSJXAd8EaSgeoWS7o9IlZ1bRMR3wC+kW7/NuAzEbHjQF5vMNXubGL2QT4bMLPy0NdUlZvT+2cP8NinAmsjYh2ApFuA84FVvWx/IXDzAb7WoIkIanfu5VUvO6jQoZiZDYks8xEskLRYUoOk1rT6Zk+GY88ANuQs16brenqNMcA5wK96ef4SSUskLamrq8vw0gdu1942Gls7mDFpdF5fx8xsuMjSj+Bakl/ra4DRwMeA/8iwX08V7NHLtm8D/q+3aqGIuD4i5kfE/KlT8ztd8sZdTQBuKDazspFprKGIWCupMiI6gB9JejDDbrXArJzlmcCmXra9gGFQLQRQu3MvADMn+4zAzMpDpg5lkkaSdCq7GtgMjM2w32JgnqS5wEaSL/v3d99I0kTgdcBFmaPOo9qdyRnBLJ8RmFmZyFI1dDFQCVwGNJL8yn93fztFRHu6z13AauDnEbFS0qWSLs3Z9J3A3RHRONDg86F2ZxPjR1UxYXTWgVnNzIpblg5lXVcNNQFfGcjBI2IRsKjbuoXdlm8EbhzIcfOpdudeZkwejeQ+BGZWHnpNBJKW03vjLhFxfF4iKrDanU1uHzCzstLXGcEBdSQrZhHBxp1NLDjcfQjMrHz01aFsX0cySYeQdBALYHFEPD8EsQ25PU3t1Le0+4zAzMpKlg5lHwMeAd4FvAd4SNJf5DuwQtjgS0fNrAxluTTm88BJEbEdQNJBwIPADfkMrBC6OpPNmORLR82sfGS5fLQWqM9Zrmf/oSNKRlcfAp8RmFk5yXJGsBF4WNJtJG0E5wOPSPosQET8ax7jG1K1O/cydmQlk8aU1cycZlbmsiSCp9Nbl9vS+5KbxzG5dHSM+xCYWVnJkgi+HhHNuSskTYmIbXmKqWDch8DMylGWNoJHJC3oWpD0bpLG4pKzMe1VbGZWTrKcEXwAuEHSfcChwEHA6/MZVCHsbmpjT7P7EJhZ+cky1tBySf8E/ITkiqEzIqI275ENsY07PQ+BmZWnfhOBpB8CRwDHA0cCd0i6NiKuy3dwQ6lrHgLPTGZm5SZLG8EK4KyIeCYi7gIWACfnN6yht6mrM5mrhsyszPSbCCLiGmC2pLPTVa3AlfkMqhA2725mZFUFB40dWehQzMyGVJaxhj4O/BL4XrpqJnBrHmMqiI27mjh0YrX7EJhZ2clSNfQp4NXAHoCIWAMcnM+gCmHz7mYOdfuAmZWhLImgJSJauxYkVdHHhDXFatOuJqZPdCIws/KTJRHcL+lvgdGS3gj8Argjv2ENrfaOTrbsaWbGpOpCh2JmNuSyJIKrgDpgOfAJkjmIv5jPoIbalvoWOgNXDZlZWcrSoawT+H56K0ldl45OdyIwszKU5Yyg5O3rQ+CqITMrQ3lNBJLOkfSkpLWSruplmzMlLZO0UtL9+YynN5t2JYOrurHYzMpRlkHnAJA0NiIaB7B9JXAd8EaSWc4WS7o9IlblbDMJ+A5wTkQ8J6kgl6Vu2tXExNEjGDsq89thZlYysnQoe5WkVcDqdPkESd/JcOxTgbURsS69/PQWktnNcr0f+HVEPAcQEVsHFP0g2by7yQ3FZla2slQNXQO8GdgOEBGPAWdk2G8G+89tXJuuy3UkMFnSfZKWSvpgTweSdImkJZKW1NXVZXjpgdm4q5lDJ7p9wMzKU6Y2gojoPll9R4bdehqroXtHtCrglcBbSJLN30s6sofXvz4i5kfE/KlTp2YJeUA27fIZgZmVryyV4hskvQoISSOBy0mrifpRC8zKWZ4JbOphm21p20OjpAeAE4CnMhx/UDS2tLO7qY3pvmLIzMpUljOCS0nGG5pB8sV9Yrrcn8XAPElz0wRyAXB7t21uA14rqUrSGOA0siWZQbN5d9eloz4jMLPylOWMQBHxgYEeOCLaJV0G3AVUAjdExEpJl6bPL4yI1ZLuBB4HOoEfRMSKgb7WS9F16airhsysXGVJBA9Kegb4L+BXEbEr68EjYhHJkBS56xZ2W/4G8I2sxxxs+3oVu7HYzMpUlolp5pGMLfRy4FFJv5V0Ud4jGyKbdjVRIZg2wYnAzMpT1quGHomIz5L0DdgB3JTXqIbQpt3NHDy+mhGVHm3DzMpTlg5lEyR9SNLvgAeBzSQJoSQkl476bMDMyleWNoLHSKam/MeI+FN+wxl6m3Y18YoZEwsdhplZwWRJBIdHRMnNSAYQEWze3cybXn5IoUMxMyuYXhOBpH+LiCuB2yW9KBFExNvzGdhQ2NPcTkt7JwePH1XoUMzMCqavM4KfpPffHIpACmFbQwsAU8Y5EZhZ+eo1EUTE0vThiRHx7dznJF0BFGTugMG0rT5JBAeNG1ngSMzMCifLNZMf6mHdhwc5joLY3tgK+IzAzMpbX20EF5LMFzBXUu4YQeNJh6Qudq4aMjPru42gq8/AFOBbOevrScYGKnrb6luQYPKYEYUOxcysYPpqI3gWeBY4fejCGVrbGlupGTOSKvcqNrMylqVn8QJJiyU1SGqV1CFpz1AEl2/b6ltcLWRmZS/LT+FrgQuBNcBo4GPAf+QzqKGyraHFVwyZWdnLOujcWqAyIjoi4kfAWfkNa2hsb2z1GYGZlb0sQ0zsTWcYWybpapIG5LH5DWtouGrIzCzbGcHFJDOMXQY0ksxD/O58BjUUmlo7aGztcNWQmZW9fs8I0quHAJqAr+Q3nKHT1Ydgqs8IzKzM9dWhbDnQ66ijEXF8XiIaIvs6k433GYGZlbe+zgjeOmRRFMC2Bg8vYWYG/XcoK1nbG7oGnHMiMLPy1m8bgaR6XqgiGgmMABojYkI+A8u3rqqhg8a6asjMylu/Vw1FxPiImJDeqkmuGLo2y8ElnSPpSUlrJV3Vw/NnStotaVl6+9LAi3BgtjW0Mr66iuoRlUP1kmZmw1KWfgT7iYhbe/pS705SJXAd8EagFlgs6faIWNVt0/+JiCFvj6hrcB8CMzPIVjX0rpzFCmA+fVxNlONUYG1ErEuPcwtwPtA9ERTE9oYWprgPgZlZpjOCt+U8bgfWk3yh92cGsCFnuRY4rYftTpf0GLAJ+FxErOy+gaRLgEsAZs+eneGl+7etoZV5B48blGOZmRWzLB3KPnKAx1ZPh+u2/ChwWEQ0SDoPuBWY10MM1wPXA8yfPz/L2Ui/tjW0sODwmsE4lJlZUctSNTQX+DQwJ3f7iHh7P7vWkgxH0WUmya/+fSJiT87jRZK+I2lKRGzrP/QD19bRya69bW4jMDMjW9XQrcAPgTuAzgEcezEwL00kG4ELSKa+3EfSIcCWiAhJp5K0QeR9GswdnqvYzGyfLImgOSL+faAHjoh2SZcBd5EMWndDRKyUdGn6/ELgPcBfSmonGcvogogYlKqfvtTVd81V7MZiM7MsieDbkv4BuBto6VoZEY/2t2NELAIWdVu3MOfxtWTskzCYtvuMwMxsnyyJ4DiSoahfzwtVQ5EuF6Vt+84InAjMzLIkgncCh0dEa76DGSr7hpdw1ZCZWaaJaR4DJuU5jiG1vbGVUVUVjBs14I7VZmYlJ8s34TTgCUmL2b+NoL/LR4etrikqpZ66OpiZlZcsieAf8h7FENve2OpqITOzVJaexfcPRSBDaYcTgZnZPv22EUiql7QnvTVL6pC0p7/9hrMdja3UeB4CMzMg2xnB+NxlSe8gGVm0aG1vbKFmjBOBmRlku2poPxFxK0Xch6CptYPmtk5qXDVkZgbkdz6CYWl7o6eoNDPLlc/5CIalrgHnJrtqyMwMyO98BMNSVyLwVUNmZoksVw3dJGlSzvJkSTfkNao86koENWM9zpCZGWRrLD4+InZ1LUTETuCkvEWUZy8kAp8RmJlBtkRQIWly14KkGrK1LQxL2xtbqaoQE6qLtghmZoMqy7fht4AHJf2S5Gqh9wL/lNeo8mhnYyuTx470OENmZqksjcU/lrSEpO+AgHdFxKq8R5Yn2xtbfemomVmOTPUj6Rd/0X7559rR2OpLR83Mcgy4Z3Gx29nY6l7FZmY5yi4RuGrIzGx/ZZUI2jo62d3U5qohM7McZZUIdu1tA9yr2MwsV14TgaRzJD0paa2kq/rY7pR0noP35DMedyYzM3uxvCUCSZXAdcC5wLHAhZKO7WW7rwN35SuWLl0jjzoRmJm9IJ9nBKcCayNiXUS0ArfQ86ilnwZ+BWzNYyyAzwjMzHqSz0QwA9iQs1ybrttH0gzgncDCvg4k6RJJSyQtqaurO+CAdjoRmJm9SD4TQU9jOHSf0ObfgL+JiI6+DhQR10fE/IiYP3Xq1AMOaLvnIjAze5F8jrxWC8zKWZ4JbOq2zXzglnTcnynAeZLa0+kwB92OxlYmVFcxorKsLpYyM+tTPhPBYmCepLnARuAC4P25G0TE3K7Hkm4EfpuvJABJIjhonOchMDPLlbdEEBHtki4juRqoErghIlZKujR9vs92gXzY0djq9gEzs27yOih/RCwCFnVb12MCiIgP5zMWSBLBrJox+X4ZM7OiUlaV5dsbW6lxQ7GZ2X7KJhFEhEceNTPrQdkkgj3N7bR3hkceNTPrpmwSwQ73ITAz61HZJQJXDZmZ7a/sEoGrhszM9lc2iWDymBGc+4pDOGRCdaFDMTMbVvLaj2A4mT+nhvlzagodhpnZsFM2ZwRmZtYzJwIzszLnRGBmVuacCMzMypwTgZlZmXMiMDMrc04EZmZlzonAzKzMKaL7fPLDm6Q64NkD3H0KsG0QwxkuSrFcpVgmKM1yuUzF4bCImNrTE0WXCF4KSUsiYn6h4xhspViuUiwTlGa5XKbi56ohM7My50RgZlbmyi0RXF/oAPKkFMtVimWC0iyXy1TkyqqNwMzMXqzczgjMzKwbJwIzszJXNolA0jmSnpS0VtJVhY7nQEiaJemPklZLWinpinR9jaR7JK1J7ycXOtaBklQp6c+Sfpsul0KZJkn6paQn0s/s9GIvl6TPpH97KyTdLKm6GMsk6QZJWyWtyFnXazkkfSH97nhS0psLE3X+lEUikFQJXAecCxwLXCjp2MJGdUDagb+KiGOABcCn0nJcBfw+IuYBv0+Xi80VwOqc5VIo07eBOyPiaOAEkvIVbbkkzQAuB+ZHxCuASuACirNMNwLndFvXYznS/7ELgJen+3wn/U4pGWWRCIBTgbURsS4iWoFbgPMLHNOARcTmiHg0fVxP8sUyg6QsN6Wb3QS8oyABHiBJM4G3AD/IWV3sZZoAnAH8ECAiWiNiF0VeLpLpbUdLqgLGAJsowjJFxAPAjm6reyvH+cAtEdESEc8Aa0m+U0pGuSSCGcCGnOXadF3RkjQHOAl4GJgWEZshSRbAwQUM7UD8G/DXQGfOumIv0+FAHfCjtMrrB5LGUsTlioiNwDeB54DNwO6IuJsiLlM3vZWj5L4/uiuXRKAe1hXtdbOSxgG/Aq6MiD2FjuelkPRWYGtELC10LIOsCjgZ+G5EnAQ0UhxVJr1K68zPB+YChwJjJV1U2KiGREl9f/SkXBJBLTArZ3kmySlt0ZE0giQJ/DQifp2u3iJpevr8dGBroeI7AK8G3i5pPUmV3esl/SfFXSZI/uZqI+LhdPmXJImhmMt1NvBMRNRFRBvwa+BVFHeZcvVWjpL5/uhNuSSCxcA8SXMljSRp+Lm9wDENmCSR1Dmvjoh/zXnqduBD6eMPAbcNdWwHKiK+EBEzI2IOyefyh4i4iCIuE0BEPA9skHRUuuoNwCqKu1zPAQskjUn/Ft9A0k5VzGXK1Vs5bgcukDRK0lxgHvBIAeLLn4goixtwHvAU8DTwd4WO5wDL8BqSU9LHgWXp7TzgIJKrHNak9zWFjvUAy3cm8Nv0cdGXCTgRWJJ+XrcCk4u9XMBXgCeAFcBPgFHFWCbgZpJ2jjaSX/wf7ascwN+l3x1PAucWOv7BvnmICTOzMlcuVUNmZtYLJwIzszLnRGBmVuacCMzMypwTgZlZmXMisH0k3Scp7xN2S7o8HY3zp/l+rUJKRx/95AD3uS8d4fLxdNTSayVNylOIua+7qL/XkfSPks7uYf2Nkp6R9JikpyT9OB2gbrBiWy9pygC2/7Cka9PHX5b0ucGKpVQ5EdigSAchy+qTwHkR8YF8xTNMTCIp60B9ICKOB44HWhiCDloRcV4kg+L1tc2XIuLeXp7+fEScABwF/Bn4Y9p58yUptVE+hysngiIjaU76a/r76bjwd0sanT637xe9pCnpsA1dv5BulXRH+svtMkmfTQdDe0hSTc5LXCTpwXS8+VPT/cem47cvTvc5P+e4v5B0B3B3D7F+Nj3OCklXpusWkgzIdrukz3TbvlLSNyUtT38Rfzpd/4b0dZencYxK16+X9M+S/iRpiaSTJd0l6WlJl6bbnCnpAUm/kbRK0kJJFelzF6bHXCHp6zlxNEj6p/QX7kOSpqXrp0r6Vfo+LJb06nT9l9O47pO0TtLl6aH+BThC0jJJ35A0PY1lWfqar+3rs45kpNy/BmZLOiF9rVslLU0/+0syxHyjpO8qmcdinaTXpbGulnRjzv7r07+Zvv6+bpT0nn5ijoi4BnieZNj3vt7nvt7/f5T0MHB6uvrzkh5Jby/r6/OwA1DoHm2+DewGzCGZl+DEdPnnwEXp4/tIxooHmAKsTx9/mGTo3PHAVGA3cGn63DUkg9d17f/99PEZwIr08T/nvMYkkh7aY9Pj1tJDT1LglcDydLtxwErgpPS59cCUHvb5S5JxlKrS5RqgmmTkxyPTdT/OiXc98Jc55Xg8p4xb0/VnAs0kyacSuAd4D8mgac+l21YBfwDeke4TwNvSx1cDX0wf/wx4Tfp4NslQHwBfBh4k6WU7BdgOjEg/qxU55fsr0l7taSzje3gP9n2GOetuBd7X9Z6k96NJevce1E/MN5KM4SSSAeP2AMeR/Ahcygt/R+vT2OfQ+9/XjcB7eoj5RetJRpT9m97e5wzv/3tzjrU+5337IC/0Pu/t8/gwcG3OZ/O5Qv/fDvfbQE7nbfh4JiKWpY+Xkvzz9uePkcxhUC9pN3BHun45SRVEl5shGa9d0gQl9cZvIhkYrquutZrkHw/gnojoPq47JMNh/CYiGgEk/Rp4LUm1QW/OBhZGRHsaw470l/AzEfFUus1NwKdIvmjghTGjlgPjcsrYrBfqvB+JiHVpHDensbUB90VEXbr+pyTJ71agFfhtuu9S4I058R0r7RuMcoKk8enj/46IFqBF0lZgWg/lWwzcoGTgwFtzPsP+5I5+ebmkd6aPZ5GMe7O9j5gB7oiIkLQc2BIRy9MyryT52+kex4H8ffUW8yn0/D5HL+tvBTpIfhDkujnn/pr0cV+fhw2AE0Fxasl53EHy6xCSX3Jd1X3VfezTmbPcyf5/B93HHAmSf+p3R8STuU9IOo1keOWe9DR0b3/Uw+v3d5zccnQvY1e5eitTb9oi/TlJ8v52HacCOD0imvYLMPki6v6ZvOh/K02uZ5BMwvMTSd+IiB/3EUdXHflxwGpJZ5J8+Z0eEXsl3ccLn3NvMefG1td7lKu3v6+BOIlkvJ7eqp/7ev+bI6Kj27ro4XFfn4cNgNsISst6kioZSKo/DsT7ACS9hmTikd3AXcCnlf6HSTopw3EeAN6hZKTKscA7gf/pZ5+7gUuVNjwrabt4ApjTVS8MXAzcP8Aynapk5NkKkvL9L8mEPq9L68UrgQszHPdu4LKuBUkn9rN9PUlVVdf2h5FUWX2fZBTZk/vaOT1z+BqwISIeByYCO9MkcDTJdKXDihKXA9OBO+n9fR7o+/++nPs/pY8H+nlYL3xGUFq+Cfxc0sUkda4HYqekB4EJwF+k675KUhXzeJoM1gNv7esgEfFo2hjZNVzvDyKir2ohSKaqPDJ9nTaS9oprJX0E+EWaIBYDCwdYpj+RNNweR5KgfhMRnZK+APyR5Nfpoojo7+qcy4HrJD1O8r/zAHBpbxtHxHZJ/6dkgvTfkdTpfz4tWwNJfXdPfiqphaTN4V5emFb1TpJE+TjJKJgP9RPvUPqGpL8nmb7yIeCsSBq7N/f2Pg/w/R+VNh5XkCQNGODnYb3z6KNW0tLqlM9FRJ+Jy6ycuWrIzKzM+YzAzKzM+YzAzKzMORGYmZU5JwIzszLnRGBmVuacCMzMytz/Bwpc40TMkmx5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = danmini_doorbell_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Danmini Doorbell')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = ecobee_thermostat_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Ecobee Thermostat')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = ennio_doorbell_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Ennio Doorbell')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = baby_monitor_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Baby Monitor')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = provision_cam1_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Provision Camera 1')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = provision_cam2_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Provision Camera 2')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = samsung_cam_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Samsung Camera')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy86wFpkAAAACXBIWXMAAAsTAAALEwEAmpwYAAApmUlEQVR4nO3deZhdVZnv8e+vpswjCYQkZBADGmaIjIoIguAEqLclioKtjXiNgIrdtO1t9dq2A07YoGlUBtGGK4gINs0gLUMblCQYMkFICEmoJJLKnEolNb73j70rOanUsKvIqVOnzu/zPOeps/fZZ+93nara79lr7bWWIgIzMytdZYUOwMzMCsuJwMysxDkRmJmVOCcCM7MS50RgZlbiKgodQHeNGTMmpkyZUugwzMyKyvz58zdGxNj2Xiu6RDBlyhTmzZtX6DDMzIqKpNUdveaqITOzEudEYGZW4pwIzMxKnBOBmVmJcyIwMytxeUsEkm6RtEHS4g5el6QfSlohaaGkE/MVi5mZdSyfVwS3Aed38voFwLT0cQXw4zzGYmZmHchbP4KIeFLSlE42uRD4eSTjYP9J0khJh0bE+nzFZP1PS0vQ2NJCU3PQ2NxCY3PQlC43tQRNzS00tQTNrY8IWtLnLQEtEekjfZ6zPnLXB+ly0NICQbKe1m2BZDHSn8mKSGPc+3qidfj3/d6z5/W969ln/d73wf7v2fOc/YeXb2/E+W4NQu8h6wtuxpTRnHlEu33CXpNCdiibALySs1ydrtsvEUi6guSqgUmTJvVKcNZzEUF9UwvbdzdSu7uJ2vomanc3saO+iV0NzdQ1NLOrsZndjc3sSp/vamxmd0Mzu5ua2d3YQn1TM/WNLTQ0t+T8bE5+NrXsOek3t/jk1JukQkdQ2q586+H9LhG09yfV7n91RNwM3AwwY8YM/+cXQGNzCxt21FOTPjbW1rNxRz2bdjawsbaezTsb2LyzgS11DWypa6ShqSXTfsvLxKDKcgZWljOwsmzPzwEV5QyoKGPowAoGVJRRVVFOVXkZAyrLqCovo6qijMpyUVGWPK8oExXlZVSVJz8rykRleRnlZaKiTJSnj7J0uUytjyQGSUhQrmS71tjKJMrLAPa+LoHSZQnK2q5LNt+zXCYh2LNN619+67Zl6dl1z+vsfT33ZxrF3vV7tst5T85n295JWz6TWzsKmQiqgcNylicC6woUiwHbdjWysqaWlzfuZNXGnazaVMfqTTtZt203G2vr260ZGDaggjHDBjB6SBWHjR7MsRNHMGpwFSMGVzJsYCXDBlQwdEAFQwcmPwdXlTO4qoJBleUMqiqnslw+OZkVWCETwf3ALEl3AacA29w+0Dsam1tYsaGWpeu2s3T9dpb9dQcvvrqDDTvq92xTJhg/chCTDxrM2eMOZtyIgYwbMZCDhw1g7LABjBk6gIOGVjGgoryAJTGzAyFviUDSncBZwBhJ1cCXgUqAiJgNPAi8E1gB1AEfy1cspe7V7buZt2oL81dv4bnqrSxeu436tOpmYGUZRxwyjLdMG8u0Q4Zy+NihTB0zhEmjB1NV4W4mZqUgn3cNzezi9QA+na/jl7JtdY3MeWkjT63YyB9XbGT1pjogOekfM2EEHzl1MsdMHMFR40cwdcyQPXXiZlaaim4YattfRLBk3XYeWfoqTy2v4blXttISMHRABae+bjSXnjKZN00dzVHjh1NZ7m/5ZrYvJ4Ii1Xryv/+5dTy4aD3VW3ZRJjjusJHMOnsaZ04bw3GHjfSJ38y65ERQZLbVNfKrea/wq3mvsHxDLZXl4s2vH8NVZ0/j7dMPYfSQqkKHaGZFxomgSLxUU8vP/udlfvPsWnY1NjNj8ii+fvHRvOuYQxk52Cd/M+s5J4I+7i9rtvDjx1/i0edfpaq8jIuOn8Blp09h+vjhhQ7NzPoJJ4I+qra+iW/+1/P84k9rGDm4ks+cPY2PnjaZMUMHFDo0M+tnnAj6oDkrNvKFexaybtsu/vaMqXz+vCMYMsC/KjPLD59d+pCWluDHT7zEdx9ZxpQxQ7jnytM5afKoQodlZv2cE0EfUVvfxDV3LeD3z7/Ke48bzzfed4yvAsysV/hM0wfs2N3I5bfOZcErW/nye6Zz+elTPBCbmfUaJ4IC27G7kctueYaF1du4ceYJXHDMoYUOycxKjBNBAdU1NO1NAh86gfOPdhIws97n8QcKpLG5hU//8lkWvLKVf5vpJGBmheMrggKICL547yL+sKyGr198tKuDzKygfEVQAD/4/XLunl/N1edM48OnTC50OGZW4pwIetlDi9dzw2PL+cBJE7nm7dMKHY6ZmRNBb1r+6g4+/6vnOO6wkfzLRUf7FlEz6xOcCHrJ9t2NXHHHfAZVVfDvl57EwErP9WtmfUOmRCBpsqS3p88HSRqW37D6n2/91wus3rSTH196IuNGDCx0OGZme3SZCCT9HXAP8O/pqonAfXmMqd/5y5ot/Mcza7j89Km8acroQodjZraPLFcEnwbOALYDRMRy4OB8BtWfNDW38E+/WcwhwwbyufOOKHQ4Zmb7yZII6iOioXVBUgUQ+Qupf7n96dUsXb+dL79nOkM9iJyZ9UFZEsETkr4IDJJ0LnA38ECWnUs6X9IySSskXdfO66Mk/UbSQknPSDq6e+H3bRtr6/n+oy9y1pFjOf/ocYUOx8ysXVkSwXVADbAI+CTwIPClrt4kqRy4CbgAmA7MlDS9zWZfBBZExLHAR4Ebsofe9/3wseXsamzmn9893beKmlmflaWuYhBwS0T8BPac4AcBdV2872RgRUSsTN93F3AhsDRnm+nANwAi4gVJUyQdEhGvdq8Yfc/Kmlr+489r+NDJk3jd2KGFDsfMrENZrggeIznxtxoE/D7D+yYAr+QsV6frcj0HvA9A0snAZJK7kvYh6QpJ8yTNq6mpyXDowvv2Q8sYUFHGVee497CZ9W1ZEsHAiKhtXUifD87wvvbqQto2Mn8TGCVpAfAZ4C9A035virg5ImZExIyxY8dmOHRhzV+9mYeW/JVPvvVwxg7zZPNm1rdlqRraKenEiHgWQNJJwK4M76sGDstZngisy90gIrYDH0v3K+Dl9FHUfvD75YwZOoBPvGVqoUMxM+tSlkRwDXC3pNaT+KHABzO8by4wTdJUYC1wCfCh3A0kjQTq0ttTPwE8mSaHorViww6eWr6Ra887gsFVvl3UzPq+Ls9UETFX0huAI0mqe16IiMYM72uSNAt4GCgnaXBeIunK9PXZwBuBn0tqJmlE/njPi9I33PrHVVRVlDHz5EmFDsXMLJOsX1nfBExJtz9BEhHx867eFBEPktxumrtuds7zp4F+05q6ra6Re59dy0XHj+egoW4bMLPi0GUikHQHcDiwAGhOVwfQZSIoNf9v3hp2NTZz+eluGzCz4pHlimAGMD0iPKxEJ5qaW7h9zmpOmTqa6eOHFzocM7PMstw+uhjw+AhdeOLFGtZu3cXHzphS6FDMzLolyxXBGGCppGeA+taVEfHevEVVhH63cD0jB1dyzhsPKXQoZmbdkiURfCXfQRS7+qZmfr/0VS44ZhyV5Z70zcyKS5bbR5/ojUCK2VMvbmRHfRPvPObQQodiZtZtWWYoO1XSXEm1khokNUsq6k5fB9qDi9YzfGAFpx8+ptChmJl1W5Z6jBuBmcBykgHnPpGuM5JqoUeXvsp5R42jqsLVQmZWfDJ1KIuIFZLKI6IZuFXSnDzHVTT+uCKpFnqXq4XMrEhlSQR1kqqABZK+DawHhuQ3rOLxnwv/yvCBFZzxelcLmVlxylKX8RGSsYJmATtJRhR9fz6DKhaNzS08uvSvnDvd1UJmVryy3DW0On26C/hqfsMpLs+u3sL23U2cO/3gQodiZtZjHSYCSb+KiL+RtIj9J5QhnWe4pD3+Yg0VZXK1kJkVtc6uCK5Of767NwIpRo8vq+GkyaMYNrCy0KGYmfVYhxXbEbE+naj+ZxGxuu2jF2Psk/66bTfPr9/OWUe6WsjMilunLZzp7aJ1kkb0UjxF44kXNwDwtjf0/TmUzcw6k+X20d3AIkmPktw1BEBEXJW3qIrA48tqGDd8IEceMqzQoZiZvSZZEsF/pg9LNTa38D/LN/KuYw9FUqHDMTN7TbLcPnp7bwRSTOav3sKO+ibOOtLVQmZW/LJMVTkN+AYwHRjYuj4iXpfHuPq0x5f5tlEz6z+ydIe9Ffgx0AS8jWSu4jvyGVRfN+eljZw4ybeNmln/kCURDIqIxwClt45+BTg7v2H1Xdt3N7J47TZOPfygQodiZnZAZEkEuyWVAcslzZJ0MZDp5nlJ50taJmmFpOvaeX2EpAckPSdpiaSPdTP+XvfMys20BJz2OicCM+sfsiSCa4DBwFXAScClwGVdvSntjHYTcAFJ+8JMSdPbbPZpYGlEHAecBXw3Hem0z3p65SaqKso4YdLIQodiZnZAZLl9tCkiaoFaoDvf2E8GVkTESgBJdwEXAktztglgmJJ7MIcCm0naIvqsOS9t4qRJoxhYWV7oUMzMDogsVwTfk/SCpK9JOqob+54AvJKzXJ2uy3Uj8EZgHbAIuDoiWtruSNIVkuZJmldTU9ONEA6sLTsbeH79dk5z+4CZ9SNdJoKIeBtJtU0NcLOkRZK+lGHf7fW0ajuK6TuABcB44HjgRknD24nh5oiYEREzxo4t3L37f355E4ATgZn1K5lmU4mIv0bED4ErSU7c/5zhbdUkk9i0mkjyzT/Xx4B7I7ECeBl4Q5aYCuHplzYxqLKc4yaOLHQoZmYHTJeJQNIbJX1F0mKSqpw5JCf1rswFpkmamjYAXwLc32abNcA56XEOAY4EVnYj/l719MpNzJgyyrORmVm/kqWx+FbgTuC8iGj7jb5DEdEkaRbwMMlUl7dExBJJV6avzwa+BtyWTn4j4B8iYmN3C9EbNtbW8+KrtVx4fNtmDjOz4pZlrKFTe7rziHgQeLDNutk5z9cB5/V0/73pmZc3A3Cq+w+YWT/jOo6M5q/ewoCKMo6Z4KkZzKx/cSLI6Nk1Wzh24gi3D5hZv+OzWga7G5tZsnY7J04aVehQzMwOuA7bCCQ9wP73/e8REe/NS0R90JJ122hobuHEyU4EZtb/dNZY/J305/uAccAv0uWZwKo8xtTnPLt6K4CvCMysX+owEUTEEwCSvhYRZ+a89ICkJ/MeWR8yf/UWJo0ezNhhAwodipnZAZeljWCspD2zkUmaCpTMHI0RwbNrtnCiRxs1s34qS4eyzwKPS2rt8TsF+GTeIupj1m7dxYYd9Zzk9gEz66eydCh7KJ23uHUMoBcioj6/YfUd81dvAeAEtw+YWT+VZayhwcAXgFkR8RwwSdK78x5ZH/GXNVsZXFXOG8YNK3QoZmZ5kXXy+gbgtHS5GviXvEXUx8xfvYXjJo6kotxdLsysf8pydjs8Ir4NNAJExC7an2ug39nd2Mzz67d7Wkoz69eyJIIGSYNIO5dJOhwoiTaC59dvp6klOO6wkYUOxcwsb7LcNfRl4CHgMEm/BM4ALs9nUH3ForXbADh2ogeaM7P+K8tdQ49KehY4laRK6Oq+OmfAgbawehtjhlYxbvjAQodiZpY3Wa4IAAYCW9Ltp0siIvp97+JF1ds4ZsIIpJJoEjGzEtVlIpD0LeCDwBKgJV0dQL9OBLsamlm+YQfvOHpcoUMxM8urLFcEFwFHllInMoCl67fREnCsJ6Ixs34uy11DK4HKfAfS1yysThqKj3FDsZn1c1muCOqABZIeI+e20Yi4Km9R9QGLqrdx8LABHOKGYjPr57IkgvvTR0lZuHabbxs1s5KQ5fbR23u6c0nnAzcA5cBPI+KbbV7/AvDhnFjeCIyNiM09PeaBsLO+iZdqannPseMLGYaZWa/obKrKX0XE30haRDtTVkbEsZ3tWFI5cBNwLsn4RHMl3R8RS3P2cT1wfbr9e4DPFjoJACxZt50IdyQzs9LQ2RXB1enPno40ejKwIiJWAki6C7gQWNrB9jOBO3t4rANqYfVWAI72HUNmVgI6m6pyffpzdQ/3PQF4JWe5GjilvQ3Toa7PB2b18FgH1JJ12xk3fKCnpjSzkpBlPoJTJc2VVCupQVKzpO0Z9t1ed9z9qphS7wH+2FG1kKQrJM2TNK+mpibDoV+blRt38vqDh+b9OGZmfUGWfgQ3klTbLAcGAZ8A/i3D+6qBw3KWJwLrOtj2EjqpFoqImyNiRkTMGDs2/9Mlr960kyljBuf9OGZmfUGm2VYiYgVQHhHNEXEr8LYMb5sLTJM0VVIVycl+v9tQJY0A3gr8NnvY+bO1roGtdY1MOWhIoUMxM+sVmTqUpSfyBZK+DawHujxLRkSTpFnAwyS3j94SEUskXZm+Pjvd9GLgkYjY2aMSHGCrNtUBMNmJwMxKRJZE8BGSE/ks4LMk1T3vz7LziHgQeLDNutltlm8Dbsuyv96welOSj6Yc5KohMysNWTqUtd41tAv4an7DKbxVG+uQ4LDRTgRmVho661DWbkeyVl11KCtWqzftZPyIQQysLC90KGZmvaKzK4KediQraqs27WSyq4XMrIR0eNdQRKxufZCMOnoccCxQ/xo6mfV5qzbVuaHYzEpKlg5lnwCeAd4HfAD4k6S/zXdghbBtVyObdza4odjMSkqWu4a+AJwQEZsAJB0EzAFuyWdghbDGt46aWQnK0qGsGtiRs7yDfccQ6jdWtd466l7FZlZCslwRrAX+LOm3JHcRXQg8I+lzABHxvTzG16tWbUwSweTRviIws9KRJRG8lD5atQ4FMezAh1NYqzbVMW74QAZV+dZRMysdWRLBtyJid+4KSWMiYmOeYiqY1b511MxKUJY2gmckndq6IOn9JI3F/c6qTXUebM7MSk6WK4IPA7dIehwYDxwEnJ3PoAqhtr6JjbX1THZDsZmVmCxjDS2S9HXgDpI7hs6MiOq8R9bLWhuKp/qKwMxKTJeJQNLPgMNJehUfATwg6caIuCnfwfWmNZuTPgST3EZgZiUmSxvBYuBtEfFyRDwMnAqcmN+wet+6rbsAmDjSicDMSkuXiSAivg9MkvT2dFUDcE0+gyqEtVt3MaSqnOGDsjSbmJn1H1nGGvo74B7g39NVE4H78hhTQazbuovxIwchqdChmJn1qixVQ58GzgC2A0TEcuDgfAZVCOu27mb8yEGFDsPMrNdlSQT1EdHQuiCpgk4mrClWrVcEZmalJksieELSF4FBks4F7gYeyG9YvWt3YzObdjYwYeTAQodiZtbrsiSC64AaYBHwSZLJ6L+Uz6B6W+sdQ74iMLNSlKVDWQvwk/TRL63bmgyl5ERgZqUoyxVBj0k6X9IySSskXdfBNmdJWiBpiaQn8hlPR1qvCCY4EZhZCcrbTfOSyoGbgHNJJreZK+n+iFias81I4EfA+RGxRlJB7kZau3UXEhwy3G0EZlZ6Ml8RSOruIDwnAysiYmV619FdJJPa5PoQcG9ErAGIiA3dPMYBsW7rLg4eNoCqirxeIJmZ9UlZOpSdLmkp8Hy6fJykH2XY9wT2ndKyOl2X6whglKTHJc2X9NEOYrhC0jxJ82pqajIcunvWbfOto2ZWurJ8Bf4+8A5gE0BEPAecmeF97XXRbdv/oAI4CXhXeoz/I+mI/d4UcXNEzIiIGWPHjs1w6O5xZzIzK2WZ6kIiou1k9c0Z3lYNHJazPBFY1842D0XEznTGsyeB47LEdKBEBGu37nJDsZmVrCyJ4BVJpwMhqUrStaTVRF2YC0yTNFVSFXAJcH+bbX4LvEVShaTBwCkZ933AbNrZQENTC+NHuKHYzEpTlruGrgRuIKnfrwYeIRl/qFMR0SRpFvAwUA7cEhFLJF2Zvj47Ip6X9BCwEGgBfhoRi3tWlJ5xZzIzK3VZEoEi4sM92XlEPEjSEzl33ew2y9cD1/dk/weCE4GZlbosVUNzJD0i6ePpff/9ytq0V7HbCMysVGWZmGYaydhCRwHPSvqdpEvzHlkvWbd1F4Mqyxk5uLLQoZiZFUTWu4aeiYjPkXQS2wzcnteoelEy/PRAT0hjZiUrS4ey4ZIuk/RfwBxgPUlC6Bc8D4GZlbosjcXPkUxN+X8j4un8htP71m7dzRsPHV7oMMzMCiZLInhdRPS7GckAGppa2Fhbz6EjfEVgZqWrw0Qg6QcRcQ1wv6T9EkFEvDefgfWGjbX1ABw8fECBIzEzK5zOrgjuSH9+pzcCKYSaHUkiGDvUicDMSleHiSAi5qdPj4+IG3Jfk3Q1UJBJZA6kPYlgmBOBmZWuLLePXtbOussPcBwFUVPrRGBm1lkbwUySiWOmSsodLG4Y6ZDUxa71iuCgoVUFjsTMrHA6ayNo7TMwBvhuzvodJIPEFb2aHfWMHFzJgIryQodiZlYwnbURrAZWA6f1Xji9a8OO3RzsaiEzK3FZehafKmmupFpJDZKaJW3vjeDyrWZHvdsHzKzkZWksvhGYCSwHBgGfAP4tn0H1lpraet86amYlL0vPYiJihaTyiGgGbpU0J89x5V1E+IrAzIxsiaAunWpygaRvkzQgD8lvWPlXW9/E7sYWJwIzK3lZqoY+QjLV5CxgJ8mE9O/PZ1C9wZ3JzMwSXV4RpHcPAewCvprfcHrP3uElPGm9mZW2zjqULQI6HHU0Io7NS0S9ZMMODzhnZgadXxG8u9eiKAAPOGdmluiqQ1m/VVNbT2W5GDHIcxWbWWnL0qFsh6Tt6WN3dzqUSTpf0jJJKyRd187rZ0naJmlB+vjnnhSiJ2p21DNm6ADKyjxXsZmVtiyNxcNylyVdRIY5iyWVAzcB5wLVwFxJ90fE0jabPhURvV4N5T4EZmaJLLeP7iMi7gPOzrDpycCKiFgZEQ3AXcCF3T1evtTscK9iMzPIcEUg6X05i2XADDq5myjHBOCVnOVq4JR2tjtN0nPAOuDaiFjSTgxXAFcATJo0KcOhu1ZTW8+xE0cckH2ZmRWzLD2L35PzvAlYRbZv9u1VvrdNIM8CkyOiVtI7gfuAafu9KeJm4GaAGTNmZElCnWpuCTbVumrIzAyytRF8rIf7ribphdxqIsm3/tx9b895/qCkH0kaExEbe3jMTDbtrKcl8BDUZmZkqxqaCnwGmJK7fUS8t4u3zgWmpe9fC1xCMuNZ7r7HAa9GREg6maTqKe+zn3l4CTOzvbJUDd0H/Ax4AGjJuuOIaJI0C3iYZKyiWyJiiaQr09dnAx8APiWpiWQIi0si4jVX/XTFicDMbK8siWB3RPywJzuPiAeBB9usm53z/EaS+Q56lccZMjPbK0siuEHSl4FHgPrWlRHxbN6iyrOa2qQYY4Z50nozsyyJ4BiSoajPZm/VUJCtL0GfVLOjnqEDKhhclWleHjOzfi3LmfBi4HVpp7B+IRlewlcDZmaQrWfxc8DIPMfRq7bUNTB6iBOBmRlkuyI4BHhB0lz2bSPo6vbRPmvzzkYmjHRDsZkZZEsEX857FL1sa10DR40fXugwzMz6hCw9i5/ojUB605a6BkYN9jwEZmaQrWfxDvaOEVQFVAI7I6Iov1Lvamhmd2MLo9xGYGYG5HE+gr5qS11y89OowU4EZmaQ3/kI+qS9icBVQ2ZmkN/5CPqkLTsbAV8RmJm1yud8BH3SnisCtxGYmQH5nY+gT3IbgZnZvrpsI5B0u6SROcujJN2S16jyqLVqaKTbCMzMgGyNxcdGxNbWhYjYApyQt4jybEtdA8MGVFBZ3u12cjOzfinL2bBM0qjWBUmjyda20CdtqWtw+4CZWY4sJ/TvAnMk3UNyt9DfAF/Pa1R5tKWu0beOmpnlyNJY/HNJ80j6Dgh4X0QszXtkebK1rsENxWZmOTJV8aQn/qI9+efavLOBw8cOLXQYZmZ9Rsm1mG6ta/QdQ2ZmOUoqETQ0tVBb3+SqITOzHCWVCLa6V7GZ2X7ymggknS9pmaQVkq7rZLs3SWqW9IF8xrOlrnWcIVcNmZm1ylsikFQO3ARcAEwHZkqa3sF23wIezlcsrTy8hJnZ/vJ5RXAysCIiVkZEA3AX7Q9W9xng18CGPMYCwJadTgRmZm3lMxFMAF7JWa5O1+0haQJwMTC7sx1JukLSPEnzampqehzQnqqhIa4aMjNrlc9EoHbWtZ3H4AfAP0REc2c7ioibI2JGRMwYO3ZsjwNy1ZCZ2f7yOWZQNXBYzvJEYF2bbWYAd0kCGAO8U1JTOgvaAbdlZwODKssZWFmej92bmRWlfCaCucA0SVOBtcAlwIdyN4iIqa3PJd0G/C5fSQA8zpCZWXvylggioknSLJK7gcqBWyJiiaQr09c7bRfIB488ama2v7wOJx0RDwIPtlnXbgKIiMvzGQukicDtA2Zm+yixnsUeZ8jMrK2SSgSbdzYw2lVDZmb7KJlE0NwSbN/dyEhXDZmZ7aNkEsG2XY1EeJwhM7O2SiYRbE6Hl3DVkJnZvkomEbQOQe2qITOzfZVMImgdZ2i0E4GZ2T5KJhGMGlzJ+UeN45DhAwodiplZn5LXDmV9yYwpo5kxZXShwzAz63NK5orAzMza50RgZlbinAjMzEqcE4GZWYlzIjAzK3FOBGZmJc6JwMysxDkRmJmVOEVEoWPoFkk1wOoevn0MsPEAhtNX9Mdy9ccyQf8sl8tUHCZHxNj2Xii6RPBaSJoXETMKHceB1h/L1R/LBP2zXC5T8XPVkJlZiXMiMDMrcaWWCG4udAB50h/L1R/LBP2zXC5TkSupNgIzM9tfqV0RmJlZG04EZmYlrmQSgaTzJS2TtELSdYWOpyckHSbpD5Kel7RE0tXp+tGSHpW0PP05qtCxdpekckl/kfS7dLk/lGmkpHskvZD+zk4r9nJJ+mz6t7dY0p2SBhZjmSTdImmDpMU56zosh6R/TM8dyyS9ozBR509JJAJJ5cBNwAXAdGCmpOmFjapHmoDPR8QbgVOBT6fluA54LCKmAY+ly8XmauD5nOX+UKYbgIci4g3AcSTlK9pySZoAXAXMiIijgXLgEoqzTLcB57dZ12450v+xS4Cj0vf8KD2n9BslkQiAk4EVEbEyIhqAu4ALCxxTt0XE+oh4Nn2+g+TEMoGkLLenm90OXFSQAHtI0kTgXcBPc1YXe5mGA2cCPwOIiIaI2EqRl4tkettBkiqAwcA6irBMEfEksLnN6o7KcSFwV0TUR8TLwAqSc0q/USqJYALwSs5ydbquaEmaApwA/Bk4JCLWQ5IsgIMLGFpP/AD4e6AlZ12xl+l1QA1wa1rl9VNJQyjickXEWuA7wBpgPbAtIh6hiMvURkfl6Hfnj7ZKJRGonXVFe9+spKHAr4FrImJ7oeN5LSS9G9gQEfMLHcsBVgGcCPw4Ik4AdlIcVSYdSuvMLwSmAuOBIZIuLWxUvaJfnT/aUyqJoBo4LGd5IsklbdGRVEmSBH4ZEfemq1+VdGj6+qHAhkLF1wNnAO+VtIqkyu5sSb+guMsEyd9cdUT8OV2+hyQxFHO53g68HBE1EdEI3AucTnGXKVdH5eg354+OlEoimAtMkzRVUhVJw8/9BY6p2ySJpM75+Yj4Xs5L9wOXpc8vA37b27H1VET8Y0RMjIgpJL+X/46ISyniMgFExF+BVyQdma46B1hKcZdrDXCqpMHp3+I5JO1UxVymXB2V437gEkkDJE0FpgHPFCC+/ImIkngA7wReBF4C/qnQ8fSwDG8muSRdCCxIH+8EDiK5y2F5+nN0oWPtYfnOAn6XPi/6MgHHA/PS39d9wKhiLxfwVeAFYDFwBzCgGMsE3EnSztFI8o3/452VA/in9NyxDLig0PEf6IeHmDAzK3GlUjVkZmYdcCIwMytxTgRmZiXOicDMrMQ5EZiZlTgngn5G0uOS8j7ptqSr0hE1f5nvYxVSOoLo/+7me96dDivxnKSlkj6Zrr9S0kcPUFzd+j1Luk3SB9qsqz0QsXQjhpMlPZmO4PlCOuzG4N6MIQtJD0na2joSbimoKHQA1ndIqoiIpoyb/2+S+6lfzmdMfcBIkrL+KMvGac/vm4GTI6Ja0gBgCkBEzM5TjH2epEOAu4FLIuLptEPa+4FhQF0vHL87f9vXkwyo98k8htSn+IqgACRNSb9N/yQd2/0RSYPS1/Z805M0Jh16AUmXS7pP0gOSXpY0S9Ln0m+ef5I0OucQl0qak44Zf3L6/iHpGOxz0/dcmLPfuyU9ADzSTqyfS/ezWNI16brZJIOq3S/ps222L5f0HUmLJC2U9Jl0/TnpcRelcQxI16+S9K+SnpY0T9KJkh6W9JKkK9Ntzkq/Sf4m/YY9W1JZ+trMdJ+LJX0rJ45aSV9Pv5X/KT0RIWmspF+nn8NcSWek67+SxvW4pJWSrkp39U3gcEkLJF0v6dA0lgXpMd/S5iMbRvIFaxNAJCNWLss5xrU5v+fvp/t6XtKbJN2rZCz8f8n5O3lB0u3pZ3lPe9+gJZ2Xfn7Ppr/Loe382XVIievT8iyS9MGcz/0JSb+S9KKkb0r6sKRn0u0O7+wzbePTwO0R8XT6uURE3BMRryq5UpiT/n3MUdobO+vfvKTDlXyLny/pKUlvSNffJul7kv4AfKuj47QVEY8BO7rzGRa9QvdoK8UHyTfEJuD4dPlXwKXp88dJxnsHGAOsSp9fTjL87TBgLLANuDJ97fskA9C1vv8n6fMzgcXp83/NOcZIkl7WQ9L9VtNOb1DgJGBRut1QYAlwQvraKmBMO+/5FMlYSBXp8mhgIMnojUek636eE+8q4FM55ViYU8YN6fqzgN0kyacceBT4AMnAZ2vSbSuA/wYuSt8TwHvS598GvpQ+/w/gzenzSSTDdQB8BZhD0lN2DMmJvDL9XS3OKd/nSXump7EMa+cz+CnJODV3Ah8GynKOcW3O7+lb6fOrScauOTQ9fjVJL9cpaTnOSLe7pc37Z6SxPgkMSdf/A/DP7cR0G/Aye3ukLwBq09fen36m5cAh6Wd6aPq5b82Jay3w1ZyYf9DZZ9rm+PcCF3bw/zCcvX8vbwd+3c2/+ceAaenzU0iGKWkt8++A8s6O00FMZ5H2ci+Fh6uGCufliFiQPp9PWn3QhT9EMg/BDknbgAfS9YuAY3O2uxOSMdclDZc0EjiPZHC3a9NtBpL80wI8GhFtx2aHZEiL30TETgBJ9wJvAf7SSYxvB2ZHehkeEZslHZeW98V0m9tJviH+IF1uHfdpETA0p4y709gBnomIlWkcd6axNQKPR0RNuv6XJMnvPqCB5CQAyed7bk5806U9A0oOlzQsff6fEVEP1EvaQHJSbGsucIuSKqD7cn6He0TEJyQdkx7r2vTYl7ezr9xyL4l0CGRJK0kGOdsKvBIRf0y3+wXJxDDfydnHqSSTLf0xLVMV8HQ7xwL4QkTc07qgvW0EbwbujIhmkoHXngDeBGwH5ubE9RJ7rxoXAW9Ln7f7maa/xyxGALdLmkaS+CpzXuv0bz69+jkduDvn+ANy3n93Wq6ujlPSnAgKpz7neTMwKH3exN4qu4GdvKclZ7mFfX+XbccNCZKhdN8faTVFK0mnkAyR3J72ht/tito5flf7yS1H2zK2lqujMnWkMdKvdiSfb+t+yoDTImLXPgEmJ5G2v5P9/j/S5HomyUQ6d0i6PiJ+3s52i4BFku4g+SZ+eTsx9rTc+4ROkshntrP/rDr7HLP8zbX7mbaxhOQKs70B6b5GcsK/WMk8G4934/hlwNaIOL6D4+b+bXd2nJLmNoK+ZxXJPwwk1R890VrH+2aSyUO2AQ8Dn1F6xpN0Qob9PAlcpGS0ySHAxcBTXbznEeBKJTNYkdbjvgBMkfT6dJuPAE90s0wnKxk9toykfP9DMinPW5W0pZQDMzPs9xFgVuuCpOO72H4HSdVE6/aTSaqsfkIyEuyJuRtLGirprJxVxwOruzhGZyZJOi19PpOk3Ln+BJzR+tmmv6sjunmMJ4EPKmnfGUtyVdWd0TWzfKY3ApelXzxat7tU0jiSb+pr09WXdyfwSObjeFnS/0r3qfQKtD09Pk5/50TQ93wH+JSkOST1vz2xJX3/bJJRFSH5NlQJLFQyYffXutpJJNNi3kZyUvgz8NOI6KxaCJL68TXpcZ4DPhQRu4GPkVy+LyL5NtfdO2ieJmm4XUzyDfs3aZXFPwJ/AJ4Dno2IroZAvgqYoaTxdSlwZWcbR8QmkmqXxZKuJ6k7XiDpLyR16ze0eYuAv1dyi+QCktE6L89cyv09T3ICXUjS3vLjNvHVpPu/M93mT8AbunmM35C0zTxH0s7y95EMo51Vl59pRLxKMsz4d9LP5nmSasbtJG0435D0R5J2iu76MPDx9O9tCR1PQ5vpOJKeIrnD6RxJ1eqHk9W35dFHrc9Lv2FfGxHvLnAovSqtvvhdJBPFm+WNrwjMzEqcrwjMzEqcrwjMzEqcE4GZWYlzIjAzK3FOBGZmJc6JwMysxP1/rYJW8cFdSJ8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = simplehome_cam1_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Simple Home Camera 1')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = simplehome_cam2_df.drop(['label','device'], axis=1).values\n",
    "X_std = StandardScaler().fit_transform(X)\n",
    "pca = PCA().fit(X_std)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Simple Home Camera 2')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = danmini_doorbell_df.drop(['label','device'], axis=1).values\n",
    "pca = PCA().fit(X)\n",
    "\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('number of components Danmini Doorbell')\n",
    "plt.ylabel('cumulative explained variance');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----------------------------------------------------\n",
    "#### Random Forest Classifier\n",
    "----------------------------------------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "| Device                                        | All Data - No Scaling - 3 Class | All Data - Standard Scaling - 3 Class | Under sampled data- No scaling - 3 class | Under sampled data- Standard scaling - 3 class | All Data - No Scaling - 11 Class | All Data - Standard Scaling - 11 Class | Under sampled data- No scaling - 11 class | Under sampled data- Standard scaling - 11 class |\n",
    "| --------------------------------------------- | ------------------------------- | ------------------------------------- | ---------------------------------------- | ---------------------------------------------- | -------------------------------- | -------------------------------------- | ----------------------------------------- | ----------------------------------------------- |\n",
    "| Danmin DoorBell                               | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 1.0                              | 0.878                                  | 1.0                                       | 0.878                                           |\n",
    "| Ecobee_Thermostat                             | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.998                            | 0.848                                  | 0.988                                     | 1.0                                             |\n",
    "| Ennio_Doorbell                                | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.992                            | 0.621                                  | 0.983                                     | 0.776                                           |\n",
    "| Philips_B120N10_Baby_Monitor                  | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.997                            | 0.886                                  | 0.989                                     | 1.0                                             |\n",
    "| Provision_PT_737E_Security_Camera             | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.993                            | 0.833                                  | 0.981                                     | 0.878                                           |\n",
    "| Provision_PT_838_Security_Camera              | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 1.0                              | 0.857                                  | 1.0                                       | 1.0                                             |\n",
    "| Samsung_SNH_1011_N_Webcam                     | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.999                            | 0.650                                  | 0.998                                     | 0.778                                           |\n",
    "| SimpleHome_XCS7_1002_WHT_Security_Camera      | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 1.0                              | 0.862                                  | 1.0                                       | 0.877                                           |\n",
    "| SimpleHome_XCS7_1003_WHT_Security_Camera      | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.993                            | 0.847                                  | 0.970                                     | 1.0                                             |"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------------------------------------\n",
    "#### Decision Tree\n",
    "------------------------------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "| Device                                        | All Data - No Scaling - 3 Class | All Data - Standard Scaling - 3 Class | Under sampled data- No scaling - 3 class | Under sampled data- Standard scaling - 3 class | All Data - No Scaling - 11 Class | All Data - Standard Scaling - 11 Class | Under sampled data- No scaling - 11 class | Under sampled data- Standard scaling - 11 class |\n",
    "| --------------------------------------------- | ------------------------------- | ------------------------------------- | ---------------------------------------- | ---------------------------------------------- | -------------------------------- | -------------------------------------- | ----------------------------------------- | ----------------------------------------------- |\n",
    "| Danmin DoorBell                               | 1.0                             | 0.999                                 | 1.0                                      | 1.0                                            | 0.865                            | 0.744                                  | 0.574                                     | 0.453                                           |\n",
    "| Ecobee_Thermostat                             | 0.997                           | 0.997                                 | 1.0                                      | 1.0                                            | 0.925                            | 0.775                                  | 0.770                                     | 0.770                                           |\n",
    "| Ennio_Doorbell                                | 0.999                           | 0.999                                 | 1.0                                      | 1.0                                            | 0.945                            | 0.574                                  | 0.983                                     | 0.776                                           |\n",
    "| Philips_B120N10_Baby_Monitor                  | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.857                            | 0.746                                  | 0.878                                     | 0.877                                           |\n",
    "| Provision_PT_737E_Security_Camera             | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.781                            | 0.621                                  | 0.859                                     | 0.756                                           |\n",
    "| Provision_PT_838_Security_Camera              | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.799                            | 0.656                                  | 0.877                                     | 0.877                                           |\n",
    "| Samsung_SNH_1011_N_Webcam                     | 1.0                             | 1.0                                   | 1.0                                      | 1.0                                            | 0.892                            | 0.611                                  | 0.998                                     | 0.778                                           |\n",
    "| SimpleHome_XCS7_1002_WHT_Security_Camera      | 0.996                           | 0.996                                 | 1.0                                      | 1.0                                            | 0.913                            | 0.775                                  | 0.648                                     | 0.648                                           |\n",
    "| SimpleHome_XCS7_1003_WHT_Security_Camera      | 0.997                           | 0.997                                 | 1.0                                      | 1.0                                            | 0.915                            | 0.770                                  | 0.847                                     | 0.877                                           |"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_doorbell_df = load_device_data(base_directory, file_extension, 'Danmini_Doorbell')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mirai     652100\n",
       "gafgyt    316650\n",
       "benign     49548\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_doorbell_df['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Danmini_Doorbell    1018298\n",
       "Name: device, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_doorbell_df['device'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "validationdata = danmini_doorbell_df.sample(frac=0.15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(152745, 117)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validationdata.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_validation = validationdata.drop(['label', 'device'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_validation = validationdata['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_doorbell_df = danmini_doorbell_df.drop(validationdata.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(865553, 117)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_doorbell_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Shape:(865553, 115), (865553,)\n",
      "                     importance\n",
      "feature                        \n",
      "H_L0.1_weight             0.077\n",
      "MI_dir_L0.01_weight       0.073\n",
      "MI_dir_L0.1_weight        0.054\n",
      "H_L0.01_weight            0.050\n",
      "MI_dir_L3_weight          0.045\n",
      "H_L1_weight               0.040\n",
      "MI_dir_L1_weight          0.039\n",
      "HH_jit_L3_weight          0.034\n",
      "MI_dir_L0.01_mean         0.034\n",
      "H_L0.01_mean              0.029\n",
      "HH_L1_weight              0.027\n",
      "MI_dir_L0.1_mean          0.026\n",
      "H_L3_weight               0.024\n",
      "H_L0.1_mean               0.024\n",
      "MI_dir_L5_weight          0.023\n",
      "HH_jit_L1_weight          0.022\n",
      "HH_L3_weight              0.022\n",
      "H_L5_weight               0.021\n",
      "HH_L5_weight              0.019\n",
      "HH_jit_L3_mean            0.019\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 12760      0      0]\n",
      " [     1  80082      0]\n",
      " [     0      0 166823]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     12760\n",
      "      gafgyt       1.00      1.00      1.00     80083\n",
      "       mirai       1.00      1.00      1.00    166823\n",
      "\n",
      "    accuracy                           1.00    259666\n",
      "   macro avg       1.00      1.00      1.00    259666\n",
      "weighted avg       1.00      1.00      1.00    259666\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "X = danmini_doorbell_df.drop(['label', 'device'], axis=1)\n",
    "y = danmini_doorbell_df['label']\n",
    "print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "   \n",
    "    \n",
    "clf = RandomForestClassifier()\n",
    "model_res = clf.fit(X_train, y_train)\n",
    "y_pred = model_res.predict(X_test)\n",
    "y_pred_prob = model_res.predict_proba(X_test)\n",
    "lr_probs = y_pred_prob[:,1]\n",
    "ac = accuracy_score(y_test, y_pred)\n",
    "importances = pd.DataFrame({'feature':X.columns,'importance':np.round(clf.feature_importances_,3)})\n",
    "importances = importances.sort_values('importance',ascending=False).set_index('feature')\n",
    "\n",
    "f1 = f1_score(y_test, y_pred, average='weighted')\n",
    "cm = confusion_matrix(y_test, y_pred)\n",
    "print(importances.head(20))\n",
    "    \n",
    "print('Random Forest: Accuracy=%.3f' % (ac))\n",
    "\n",
    "print('Random Forest: f1-score=%.3f' % (f1))\n",
    "print('Random Forest: Confusion Matrix', cm)\n",
    "print('Random Forest : Classification Report',classification_report(y_test, y_pred))\n",
    "\n",
    "model_name = f'test_without_scaling_unbalanced_model.pkl'\n",
    "    \n",
    "best_model = clf\n",
    "best_model.version = 1.0\n",
    "best_model.pandas_version = pd.__version__\n",
    "best_model.numpy_version = np.__version__\n",
    "best_model.sklearn_version = sklearn_version\n",
    "best_model.build_datetime = datetime.now()\n",
    "    \n",
    "modelpath = 'models'\n",
    "if not os.path.exists(modelpath):\n",
    "    os.mkdir(modelpath)\n",
    "iotmodel_path = os.path.join(modelpath, model_name)\n",
    "if not os.path.exists(iotmodel_path):\n",
    "    with open(iotmodel_path, 'wb') as f:\n",
    "        pickle.dump(best_model, f)\n",
    "        \n",
    "print('---------------------- Done ---------------------')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This isn't exactly production-grade, but a quick check for development\n",
    "# These checks can save some head-scratching in development when moving from\n",
    "# one python environment to another, for example\n",
    "expected_model_version = 1.0\n",
    "model_path = 'models/test_without_scaling_unbalanced_model.pkl'\n",
    "if os.path.exists(model_path):\n",
    "    with open(model_path, 'rb') as f:\n",
    "        model = pickle.load(f)\n",
    "    if model.version != expected_model_version:\n",
    "        print(\"Expected model version doesn't match version loaded\")\n",
    "    if model.sklearn_version != sklearn_version:\n",
    "        print(\"Warning: model created under different sklearn version\")\n",
    "else:\n",
    "    print(\"Expected model not found\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_validate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                     importance\n",
      "feature                        \n",
      "H_L0.1_weight             0.077\n",
      "MI_dir_L0.01_weight       0.073\n",
      "MI_dir_L0.1_weight        0.054\n",
      "H_L0.01_weight            0.050\n",
      "MI_dir_L3_weight          0.045\n",
      "H_L1_weight               0.040\n",
      "MI_dir_L1_weight          0.039\n",
      "HH_jit_L3_weight          0.034\n",
      "MI_dir_L0.01_mean         0.034\n",
      "H_L0.01_mean              0.029\n",
      "HH_L1_weight              0.027\n",
      "MI_dir_L0.1_mean          0.026\n",
      "H_L3_weight               0.024\n",
      "H_L0.1_mean               0.024\n",
      "MI_dir_L5_weight          0.023\n",
      "HH_jit_L1_weight          0.022\n",
      "HH_L3_weight              0.022\n",
      "H_L5_weight               0.021\n",
      "HH_L5_weight              0.019\n",
      "HH_jit_L3_mean            0.019\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 7471     0     0]\n",
      " [    1 47531     0]\n",
      " [    0     0 97742]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00      7471\n",
      "      gafgyt       1.00      1.00      1.00     47532\n",
      "       mirai       1.00      1.00      1.00     97742\n",
      "\n",
      "    accuracy                           1.00    152745\n",
      "   macro avg       1.00      1.00      1.00    152745\n",
      "weighted avg       1.00      1.00      1.00    152745\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "y_pred = model.predict(X_validation)\n",
    "y_pred_prob = model_res.predict_proba(X_validation)\n",
    "lr_probs = y_pred_prob[:,1]\n",
    "ac = accuracy_score(y_validation, y_pred)\n",
    "importances = pd.DataFrame({'feature':X.columns,'importance':np.round(clf.feature_importances_,3)})\n",
    "importances = importances.sort_values('importance',ascending=False).set_index('feature')\n",
    "\n",
    "f1 = f1_score(y_validation, y_pred, average='weighted')\n",
    "cm = confusion_matrix(y_validation, y_pred)\n",
    "print(importances.head(20))\n",
    "    \n",
    "print('Random Forest: Accuracy=%.3f' % (ac))\n",
    "\n",
    "print('Random Forest: f1-score=%.3f' % (f1))\n",
    "print('Random Forest: Confusion Matrix', cm)\n",
    "print('Random Forest : Classification Report',classification_report(y_validation, y_pred))\n",
    "\n",
    "print('---------------------- Done ---------------------')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load All data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = load_all_data_class(base_directory, file_extension)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1059391, 117)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def rf_classifier(data, scaling=False,):\n",
    "    X = data.drop(['label', 'device'], axis=1)\n",
    "    y = data['label']\n",
    "    print(f'Original Shape:{X.shape}, {y.shape}')\n",
    "    if scaling == False:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        model_name = f'generic_without_scaling_unbalanced_model.pkl'\n",
    "    else:\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=47)\n",
    "        scaler = StandardScaler()\n",
    "        scaler.fit(X_train)\n",
    "        X_train = scaler.transform(X_train)\n",
    "        X_test = scaler.transform(X_test)\n",
    "        model_name = f'generic_with_scaling_unbalanced_model.pkl'\n",
    "    \n",
    "    clf = RandomForestClassifier()\n",
    "    model_res = clf.fit(X_train, y_train)\n",
    "    y_pred = model_res.predict(X_test)\n",
    "    y_pred_prob = model_res.predict_proba(X_test)\n",
    "    lr_probs = y_pred_prob[:,1]\n",
    "    ac = accuracy_score(y_test, y_pred)\n",
    "    importances = pd.DataFrame({'feature':X.columns,'importance':np.round(clf.feature_importances_,3)})\n",
    "    importances = importances.sort_values('importance',ascending=False).set_index('feature')\n",
    "\n",
    "    f1 = f1_score(y_test, y_pred, average='weighted')\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "    print(importances.head(20))\n",
    "    \n",
    "    print('Random Forest: Accuracy=%.3f' % (ac))\n",
    "\n",
    "    print('Random Forest: f1-score=%.3f' % (f1))\n",
    "    print('Random Forest: Confusion Matrix', cm)\n",
    "    print('Random Forest : Classification Report',classification_report(y_test, y_pred))\n",
    "    \n",
    "    best_model = clf\n",
    "    best_model.version = 1.0\n",
    "    best_model.pandas_version = pd.__version__\n",
    "    best_model.numpy_version = np.__version__\n",
    "    best_model.sklearn_version = sklearn_version\n",
    "    #best_model.X_columns = [col for col in X_train.columns]\n",
    "    best_model.build_datetime = datetime.now()\n",
    "    \n",
    "    modelpath = 'models'\n",
    "    if not os.path.exists(modelpath):\n",
    "        os.mkdir(modelpath)\n",
    "    iotmodel_path = os.path.join(modelpath, model_name)\n",
    "    if not os.path.exists(iotmodel_path):\n",
    "        with open(iotmodel_path, 'wb') as f:\n",
    "            pickle.dump(best_model, f)\n",
    "    \n",
    "    print('---------------------- Done ---------------------')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Shape:(1059391, 115), (1059391,)\n",
      "                     importance\n",
      "feature                        \n",
      "MI_dir_L1_weight          0.055\n",
      "H_L0.1_weight             0.052\n",
      "MI_dir_L0.1_weight        0.051\n",
      "MI_dir_L0.01_weight       0.045\n",
      "H_L5_variance             0.036\n",
      "H_L1_weight               0.033\n",
      "MI_dir_L3_weight          0.030\n",
      "H_L0.01_weight            0.026\n",
      "MI_dir_L0.01_mean         0.026\n",
      "HH_jit_L3_weight          0.025\n",
      "HH_jit_L0.1_mean          0.024\n",
      "HH_L0.01_magnitude        0.024\n",
      "HH_jit_L1_weight          0.024\n",
      "H_L0.1_mean               0.022\n",
      "HH_L1_weight              0.022\n",
      "MI_dir_L0.1_mean          0.021\n",
      "HH_L0.1_magnitude         0.021\n",
      "MI_dir_L1_mean            0.019\n",
      "HH_jit_L5_mean            0.019\n",
      "H_L0.01_variance          0.019\n",
      "Random Forest: Accuracy=1.000\n",
      "Random Forest: f1-score=1.000\n",
      "Random Forest: Confusion Matrix [[ 25003      1      0]\n",
      " [     2 127683      0]\n",
      " [     1      0 165128]]\n",
      "Random Forest : Classification Report               precision    recall  f1-score   support\n",
      "\n",
      "      benign       1.00      1.00      1.00     25004\n",
      "      gafgyt       1.00      1.00      1.00    127685\n",
      "       mirai       1.00      1.00      1.00    165129\n",
      "\n",
      "    accuracy                           1.00    317818\n",
      "   macro avg       1.00      1.00      1.00    317818\n",
      "weighted avg       1.00      1.00      1.00    317818\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "rf_classifier(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_multi = load_all_data_multi_class(base_directory, file_extension)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1059391, 117)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_multi.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Shape:(1059391, 115), (1059391,)\n",
      "                       importance\n",
      "feature                          \n",
      "MI_dir_L0.01_mean           0.051\n",
      "HH_jit_L5_mean              0.048\n",
      "HH_jit_L0.1_mean            0.045\n",
      "HH_jit_L0.01_mean           0.043\n",
      "HH_jit_L3_mean              0.042\n",
      "H_L0.01_mean                0.042\n",
      "MI_dir_L0.01_variance       0.036\n",
      "H_L0.01_variance            0.034\n",
      "HH_jit_L1_mean              0.033\n",
      "MI_dir_L0.1_mean            0.033\n",
      "MI_dir_L0.01_weight         0.032\n",
      "H_L0.01_weight              0.029\n",
      "H_L0.1_mean                 0.029\n",
      "H_L0.1_weight               0.026\n",
      "MI_dir_L1_mean              0.024\n",
      "MI_dir_L0.1_weight          0.022\n",
      "HH_L0.01_magnitude          0.020\n",
      "H_L1_mean                   0.020\n",
      "MI_dir_L0.1_variance        0.019\n",
      "HH_L0.1_magnitude           0.017\n",
      "Random Forest: Accuracy=0.997\n",
      "Random Forest: f1-score=0.997\n",
      "Random Forest: Confusion Matrix [[25003     0     0     0     0     0     0     1     0     0     0]\n",
      " [    0 23161     2     0     0     0     0     0     0     0     0]\n",
      " [    0     1 11687     1     0     0     0     0     0     0     0]\n",
      " [    0     0     0 11559     3     0     0     0     0     0     0]\n",
      " [    0     1     0     0 37780   876     0     0     0     0     0]\n",
      " [    3     0     0     0     0 42611     0     0     0     0     0]\n",
      " [    0     0     0     0     0     0 29202     0     0     8     0]\n",
      " [    0     0     0     1     0     0     0 24287     0     0     0]\n",
      " [    0     0     0     0     0     0     0     1 32881     0     0]\n",
      " [    0     0     0     0     0     0     0     1     0 55181     0]\n",
      " [    0     0     0     0     0     0     0     0     0     1 23566]]\n",
      "Random Forest : Classification Report                 precision    recall  f1-score   support\n",
      "\n",
      "        benign       1.00      1.00      1.00     25004\n",
      "  gafgyt_combo       1.00      1.00      1.00     23163\n",
      "   gafgyt_junk       1.00      1.00      1.00     11689\n",
      "   gafgyt_scan       1.00      1.00      1.00     11562\n",
      "    gafgyt_tcp       1.00      0.98      0.99     38657\n",
      "    gafgyt_udp       0.98      1.00      0.99     42614\n",
      "     mirai_ack       1.00      1.00      1.00     29210\n",
      "    mirai_scan       1.00      1.00      1.00     24288\n",
      "     mirai_syn       1.00      1.00      1.00     32882\n",
      "     mirai_udp       1.00      1.00      1.00     55182\n",
      "mirai_udpplain       1.00      1.00      1.00     23567\n",
      "\n",
      "      accuracy                           1.00    317818\n",
      "     macro avg       1.00      1.00      1.00    317818\n",
      "  weighted avg       1.00      1.00      1.00    317818\n",
      "\n",
      "---------------------- Done ---------------------\n"
     ]
    }
   ],
   "source": [
    "rf_classifier(df_multi)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Isolation forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MI_dir_L5_weight</th>\n",
       "      <th>MI_dir_L5_mean</th>\n",
       "      <th>MI_dir_L5_variance</th>\n",
       "      <th>MI_dir_L3_weight</th>\n",
       "      <th>MI_dir_L3_mean</th>\n",
       "      <th>MI_dir_L3_variance</th>\n",
       "      <th>MI_dir_L1_weight</th>\n",
       "      <th>MI_dir_L1_mean</th>\n",
       "      <th>MI_dir_L1_variance</th>\n",
       "      <th>MI_dir_L0.1_weight</th>\n",
       "      <th>...</th>\n",
       "      <th>HpHp_L0.1_pcc</th>\n",
       "      <th>HpHp_L0.01_weight</th>\n",
       "      <th>HpHp_L0.01_mean</th>\n",
       "      <th>HpHp_L0.01_std</th>\n",
       "      <th>HpHp_L0.01_magnitude</th>\n",
       "      <th>HpHp_L0.01_radius</th>\n",
       "      <th>HpHp_L0.01_covariance</th>\n",
       "      <th>HpHp_L0.01_pcc</th>\n",
       "      <th>label</th>\n",
       "      <th>device</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>679906</th>\n",
       "      <td>98.543930</td>\n",
       "      <td>72.628513</td>\n",
       "      <td>17.333453</td>\n",
       "      <td>185.268567</td>\n",
       "      <td>71.396277</td>\n",
       "      <td>29.691762</td>\n",
       "      <td>693.295171</td>\n",
       "      <td>69.640279</td>\n",
       "      <td>42.042280</td>\n",
       "      <td>5978.075614</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>387201</th>\n",
       "      <td>103.500850</td>\n",
       "      <td>539.620024</td>\n",
       "      <td>12652.364766</td>\n",
       "      <td>178.796090</td>\n",
       "      <td>500.801396</td>\n",
       "      <td>28739.635401</td>\n",
       "      <td>587.802017</td>\n",
       "      <td>433.067304</td>\n",
       "      <td>49592.395823</td>\n",
       "      <td>6247.587046</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>566.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>566.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>457838</th>\n",
       "      <td>118.257312</td>\n",
       "      <td>392.003976</td>\n",
       "      <td>57767.371752</td>\n",
       "      <td>195.147093</td>\n",
       "      <td>411.205447</td>\n",
       "      <td>54364.690249</td>\n",
       "      <td>614.207341</td>\n",
       "      <td>403.043284</td>\n",
       "      <td>55901.169273</td>\n",
       "      <td>6320.655660</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>778096</th>\n",
       "      <td>115.213824</td>\n",
       "      <td>493.425916</td>\n",
       "      <td>26254.359914</td>\n",
       "      <td>202.983827</td>\n",
       "      <td>453.348555</td>\n",
       "      <td>39590.687161</td>\n",
       "      <td>648.891699</td>\n",
       "      <td>408.260002</td>\n",
       "      <td>50750.155681</td>\n",
       "      <td>6470.044640</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>554.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>554.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>870930</th>\n",
       "      <td>177.247175</td>\n",
       "      <td>216.339547</td>\n",
       "      <td>52789.682344</td>\n",
       "      <td>291.235598</td>\n",
       "      <td>267.977544</td>\n",
       "      <td>59486.247933</td>\n",
       "      <td>818.717531</td>\n",
       "      <td>333.367530</td>\n",
       "      <td>60313.728972</td>\n",
       "      <td>6531.915136</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>412611</th>\n",
       "      <td>97.134946</td>\n",
       "      <td>510.047829</td>\n",
       "      <td>25181.153178</td>\n",
       "      <td>177.765565</td>\n",
       "      <td>467.191133</td>\n",
       "      <td>40234.094399</td>\n",
       "      <td>591.954645</td>\n",
       "      <td>415.695149</td>\n",
       "      <td>53462.641898</td>\n",
       "      <td>6245.765590</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>566.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>566.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52871</th>\n",
       "      <td>167.595412</td>\n",
       "      <td>75.469569</td>\n",
       "      <td>356.264983</td>\n",
       "      <td>277.376717</td>\n",
       "      <td>75.216262</td>\n",
       "      <td>295.126045</td>\n",
       "      <td>806.375333</td>\n",
       "      <td>74.577952</td>\n",
       "      <td>140.044481</td>\n",
       "      <td>3122.486274</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>gafgyt</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1118275</th>\n",
       "      <td>126.734576</td>\n",
       "      <td>395.698749</td>\n",
       "      <td>57169.916851</td>\n",
       "      <td>215.037980</td>\n",
       "      <td>377.976731</td>\n",
       "      <td>59787.024143</td>\n",
       "      <td>636.035233</td>\n",
       "      <td>379.640473</td>\n",
       "      <td>59567.572623</td>\n",
       "      <td>6243.812470</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>566.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>566.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>640439</th>\n",
       "      <td>108.592844</td>\n",
       "      <td>73.260005</td>\n",
       "      <td>9.812340</td>\n",
       "      <td>189.688092</td>\n",
       "      <td>72.234265</td>\n",
       "      <td>21.602467</td>\n",
       "      <td>620.697463</td>\n",
       "      <td>70.477573</td>\n",
       "      <td>36.906827</td>\n",
       "      <td>6618.485410</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>mirai</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>194863</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>gafgyt</td>\n",
       "      <td>Danmini_Doorbell</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 117 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         MI_dir_L5_weight  MI_dir_L5_mean  MI_dir_L5_variance  \\\n",
       "679906          98.543930       72.628513           17.333453   \n",
       "387201         103.500850      539.620024        12652.364766   \n",
       "457838         118.257312      392.003976        57767.371752   \n",
       "778096         115.213824      493.425916        26254.359914   \n",
       "870930         177.247175      216.339547        52789.682344   \n",
       "412611          97.134946      510.047829        25181.153178   \n",
       "52871          167.595412       75.469569          356.264983   \n",
       "1118275        126.734576      395.698749        57169.916851   \n",
       "640439         108.592844       73.260005            9.812340   \n",
       "194863           1.000000       60.000000            0.000000   \n",
       "\n",
       "         MI_dir_L3_weight  MI_dir_L3_mean  MI_dir_L3_variance  \\\n",
       "679906         185.268567       71.396277           29.691762   \n",
       "387201         178.796090      500.801396        28739.635401   \n",
       "457838         195.147093      411.205447        54364.690249   \n",
       "778096         202.983827      453.348555        39590.687161   \n",
       "870930         291.235598      267.977544        59486.247933   \n",
       "412611         177.765565      467.191133        40234.094399   \n",
       "52871          277.376717       75.216262          295.126045   \n",
       "1118275        215.037980      377.976731        59787.024143   \n",
       "640439         189.688092       72.234265           21.602467   \n",
       "194863           1.000000       60.000000            0.000000   \n",
       "\n",
       "         MI_dir_L1_weight  MI_dir_L1_mean  MI_dir_L1_variance  \\\n",
       "679906         693.295171       69.640279           42.042280   \n",
       "387201         587.802017      433.067304        49592.395823   \n",
       "457838         614.207341      403.043284        55901.169273   \n",
       "778096         648.891699      408.260002        50750.155681   \n",
       "870930         818.717531      333.367530        60313.728972   \n",
       "412611         591.954645      415.695149        53462.641898   \n",
       "52871          806.375333       74.577952          140.044481   \n",
       "1118275        636.035233      379.640473        59567.572623   \n",
       "640439         620.697463       70.477573           36.906827   \n",
       "194863           1.000000       60.000000            0.000000   \n",
       "\n",
       "         MI_dir_L0.1_weight  ...  HpHp_L0.1_pcc  HpHp_L0.01_weight  \\\n",
       "679906          5978.075614  ...            0.0                1.0   \n",
       "387201          6247.587046  ...            0.0                1.0   \n",
       "457838          6320.655660  ...            0.0                1.0   \n",
       "778096          6470.044640  ...            0.0                1.0   \n",
       "870930          6531.915136  ...            0.0                1.0   \n",
       "412611          6245.765590  ...            0.0                1.0   \n",
       "52871           3122.486274  ...            0.0                1.0   \n",
       "1118275         6243.812470  ...            0.0                1.0   \n",
       "640439          6618.485410  ...            0.0                1.0   \n",
       "194863             1.000000  ...            0.0                1.0   \n",
       "\n",
       "         HpHp_L0.01_mean  HpHp_L0.01_std  HpHp_L0.01_magnitude  \\\n",
       "679906              74.0             0.0                  74.0   \n",
       "387201             566.0             0.0                 566.0   \n",
       "457838              60.0             0.0                  60.0   \n",
       "778096             554.0             0.0                 554.0   \n",
       "870930              60.0             0.0                  60.0   \n",
       "412611             566.0             0.0                 566.0   \n",
       "52871               74.0             0.0                  74.0   \n",
       "1118275            566.0             0.0                 566.0   \n",
       "640439              74.0             0.0                  74.0   \n",
       "194863              60.0             0.0                  60.0   \n",
       "\n",
       "         HpHp_L0.01_radius  HpHp_L0.01_covariance  HpHp_L0.01_pcc   label  \\\n",
       "679906                 0.0                    0.0             0.0   mirai   \n",
       "387201                 0.0                    0.0             0.0   mirai   \n",
       "457838                 0.0                    0.0             0.0   mirai   \n",
       "778096                 0.0                    0.0             0.0   mirai   \n",
       "870930                 0.0                    0.0             0.0   mirai   \n",
       "412611                 0.0                    0.0             0.0   mirai   \n",
       "52871                  0.0                    0.0             0.0  gafgyt   \n",
       "1118275                0.0                    0.0             0.0   mirai   \n",
       "640439                 0.0                    0.0             0.0   mirai   \n",
       "194863                 0.0                    0.0             0.0  gafgyt   \n",
       "\n",
       "                   device  \n",
       "679906   Danmini_Doorbell  \n",
       "387201   Danmini_Doorbell  \n",
       "457838   Danmini_Doorbell  \n",
       "778096   Danmini_Doorbell  \n",
       "870930   Danmini_Doorbell  \n",
       "412611   Danmini_Doorbell  \n",
       "52871    Danmini_Doorbell  \n",
       "1118275  Danmini_Doorbell  \n",
       "640439   Danmini_Doorbell  \n",
       "194863   Danmini_Doorbell  \n",
       "\n",
       "[10 rows x 117 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_doorbell_df.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1120493, 117)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_doorbell_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_doorbell_df['label'] = danmini_doorbell_df['label'].replace(['benign','mirai', 'gafgyt'],[0, 1, 2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_benign_df = danmini_doorbell_df[danmini_doorbell_df['label'] == 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(49548, 117)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_benign_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_benign_test = danmini_benign_df.sample(frac=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14864, 117)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_benign_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_train_data = danmini_benign_df.drop(danmini_benign_test.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(34684, 117)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_train_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_attack_test = danmini_doorbell_df[danmini_doorbell_df['label'] == 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(754295, 117)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_attack_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "danmini_attack_test_sample = danmini_attack_test.sample(frac=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(226288, 117)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "danmini_attack_test_sample.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "doorbell_test = danmini_benign_test.append(danmini_attack_test_sample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(241152, 117)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doorbell_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import IsolationForest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = danmini_train_data.drop(['label', 'device'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = danmini_train_data['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_X = doorbell_test.drop(['label', 'device'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_y = doorbell_test['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6387      0\n",
       "32280     0\n",
       "27082     0\n",
       "28380     0\n",
       "34044     0\n",
       "         ..\n",
       "694928    1\n",
       "664033    1\n",
       "732054    1\n",
       "607198    1\n",
       "991396    1\n",
       "Name: label, Length: 241152, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6387     0\n",
       "32280    0\n",
       "27082    0\n",
       "28380    0\n",
       "34044    0\n",
       "        ..\n",
       "12317    0\n",
       "15318    0\n",
       "48159    0\n",
       "13906    0\n",
       "20964    0\n",
       "Name: label, Length: 14864, dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_y[test_y == 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "446419     1\n",
       "1051604    1\n",
       "714451     1\n",
       "491703     1\n",
       "764881     1\n",
       "          ..\n",
       "694928     1\n",
       "664033     1\n",
       "732054     1\n",
       "607198     1\n",
       "991396     1\n",
       "Name: label, Length: 226288, dtype: int64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_y[test_y == 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "IsolationForest(contamination=0.5)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = IsolationForest(contamination=0.5)\n",
    "model.fit(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "yhat = model.predict(test_X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1,  1, -1, ..., -1, -1, -1])"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "yhat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6387      0\n",
       "32280     0\n",
       "27082     0\n",
       "28380     0\n",
       "34044     0\n",
       "         ..\n",
       "694928    1\n",
       "664033    1\n",
       "732054    1\n",
       "607198    1\n",
       "991396    1\n",
       "Name: label, Length: 241152, dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "d:\\study\\ds\\venv\\lib\\site-packages\\ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n",
      "d:\\study\\ds\\venv\\lib\\site-packages\\ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "test_y[test_y == 1] = -1\n",
    "test_y[test_y == 0] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.97      1.00      0.98    226288\n",
      "           1       1.00      0.50      0.67     14864\n",
      "\n",
      "    accuracy                           0.97    241152\n",
      "   macro avg       0.98      0.75      0.83    241152\n",
      "weighted avg       0.97      0.97      0.96    241152\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_y, yhat))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import cohen_kappa_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6548804329954762\n"
     ]
    }
   ],
   "source": [
    "print(cohen_kappa_score(test_y, yhat))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import matthews_corrcoef"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6977511518369802\n"
     ]
    }
   ],
   "source": [
    "print(matthews_corrcoef(test_y, yhat))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Elliptc envelope"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.covariance import EllipticEnvelope"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = EllipticEnvelope(contamination=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:647: UserWarning: The covariance matrix associated to your dataset is not full rank\n",
      "  warnings.warn(\"The covariance matrix associated to your dataset \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2298.333611996246418 > -2320.111999216293498). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2279.008663684981912 > -2380.198361103761727). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2154.095952984273936 > -2959.533111150696641). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2139.060486463805773 > -2759.846402743711678). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2619.688506360878364 > -2706.866063148957437). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2208.691302222795457 > -2555.492239498182698). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2208.691302222795457 > -2680.157988465698509). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2188.427456680422893 > -2441.836102597757417). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2188.427456680422893 > -2281.271060402792955). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2251.422938175548097 > -3563.593651223261531). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\numpy\\linalg\\linalg.py:2099: RuntimeWarning: overflow encountered in slogdet\n",
      "  sign, logdet = _umath_linalg.slogdet(a, signature=signature)\n",
      "e:\\springboard\\venv\\lib\\site-packages\\numpy\\linalg\\linalg.py:2099: RuntimeWarning: invalid value encountered in slogdet\n",
      "  sign, logdet = _umath_linalg.slogdet(a, signature=signature)\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2155.388258869266792 > -2445.351203003979208). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2300.928914015641567 > -2924.570534584742290). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2801.326222977541420 > -3109.657351286688936). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2736.875302342397845 > -3835.604833278567185). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2195.978726287935388 > -3442.665984911750911). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2195.978726287935388 > -3069.219755907148283). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2537.907522765342492 > -2858.057616611743015). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-1366.324872422065710 > -3378.927174117573031). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-2169.518582630910714 > -2334.989784419445641). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n",
      "e:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py:166: RuntimeWarning: Determinant has increased; this should not happen: log(det) > log(previous_det) (-1461.101088013806930 > -2781.735902614092538). You may want to try with a higher value of support_fraction (current value: 0.505).\n",
      "  warnings.warn(\"Determinant has increased; this should not happen: \"\n"
     ]
    },
    {
     "ename": "UnboundLocalError",
     "evalue": "local variable 'precision' referenced before assignment",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mUnboundLocalError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-38-0c199aedc904>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32me:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_elliptic_envelope.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, X, y)\u001b[0m\n\u001b[0;32m    143\u001b[0m             \u001b[0mNot\u001b[0m \u001b[0mused\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpresent\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mAPI\u001b[0m \u001b[0mconsistency\u001b[0m \u001b[0mby\u001b[0m \u001b[0mconvention\u001b[0m\u001b[1;33m.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    144\u001b[0m         \"\"\"\n\u001b[1;32m--> 145\u001b[1;33m         \u001b[0msuper\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    146\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0moffset_\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpercentile\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m-\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdist_\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m100.\u001b[0m \u001b[1;33m*\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcontamination\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    147\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32me:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, X, y)\u001b[0m\n\u001b[0;32m    648\u001b[0m                           \"is not full rank\")\n\u001b[0;32m    649\u001b[0m         \u001b[1;31m# compute and store raw estimates\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 650\u001b[1;33m         raw_location, raw_covariance, raw_support, raw_dist = fast_mcd(\n\u001b[0m\u001b[0;32m    651\u001b[0m             \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msupport_fraction\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msupport_fraction\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    652\u001b[0m             \u001b[0mcov_computation_method\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_nonrobust_covariance\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32me:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py\u001b[0m in \u001b[0;36mfast_mcd\u001b[1;34m(X, support_fraction, cov_computation_method, random_state)\u001b[0m\n\u001b[0;32m    444\u001b[0m             \u001b[0mhigh_bound\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlow_bound\u001b[0m \u001b[1;33m+\u001b[0m \u001b[0mn_samples_subsets\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    445\u001b[0m             \u001b[0mcurrent_subset\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0msamples_shuffle\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mlow_bound\u001b[0m\u001b[1;33m:\u001b[0m\u001b[0mhigh_bound\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 446\u001b[1;33m             best_locations_sub, best_covariances_sub, _, _ = select_candidates(\n\u001b[0m\u001b[0;32m    447\u001b[0m                 \u001b[0mcurrent_subset\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mh_subset\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mn_trials\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    448\u001b[0m                 \u001b[0mselect\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mn_best_sub\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mn_iter\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m2\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32me:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py\u001b[0m in \u001b[0;36mselect_candidates\u001b[1;34m(X, n_support, n_trials, select, n_iter, verbose, cov_computation_method, random_state)\u001b[0m\n\u001b[0;32m    283\u001b[0m         \u001b[1;32mfor\u001b[0m \u001b[0mj\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mn_trials\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    284\u001b[0m             all_estimates.append(\n\u001b[1;32m--> 285\u001b[1;33m                 _c_step(\n\u001b[0m\u001b[0;32m    286\u001b[0m                     \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mn_support\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mremaining_iterations\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mn_iter\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mverbose\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    287\u001b[0m                     \u001b[0mcov_computation_method\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcov_computation_method\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32me:\\springboard\\venv\\lib\\site-packages\\sklearn\\covariance\\_robust_covariance.py\u001b[0m in \u001b[0;36m_c_step\u001b[1;34m(X, n_support, random_state, remaining_iterations, initial_estimates, verbose, cov_computation_method)\u001b[0m\n\u001b[0;32m    151\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    152\u001b[0m     \u001b[0mprevious_dist\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mdist\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 153\u001b[1;33m     \u001b[0mdist\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mlocation\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprecision\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m*\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mlocation\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msum\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    154\u001b[0m     \u001b[1;31m# Check if best fit already found (det => 0, logdet => -inf)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    155\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0misinf\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdet\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mUnboundLocalError\u001b[0m: local variable 'precision' referenced before assignment"
     ]
    }
   ],
   "source": [
    "model.fit(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import svm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OneClassSVM(degree=2, kernel='poly')"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = svm.OneClassSVM(kernel=\"poly\", degree=2)\n",
    "clf.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "yhat = clf.predict(test_X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.94      1.00      0.97    226288\n",
      "           1       0.70      0.02      0.04     14864\n",
      "\n",
      "    accuracy                           0.94    241152\n",
      "   macro avg       0.82      0.51      0.50    241152\n",
      "weighted avg       0.92      0.94      0.91    241152\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_y, yhat))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.03215740702708003\n"
     ]
    }
   ],
   "source": [
    "print(cohen_kappa_score(test_y, yhat))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.10582324581532093\n"
     ]
    }
   ],
   "source": [
    "print(matthews_corrcoef(test_y, yhat))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.neighbors import LocalOutlierFactor\n",
    "from numpy import vstack"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = LocalOutlierFactor(n_neighbors=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "composite = vstack((X, test_X))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "yhat = model.fit_predict(composite)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "result = yhat[len(X):]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.87      0.10      0.18    226288\n",
      "           1       0.05      0.79      0.10     14864\n",
      "\n",
      "    accuracy                           0.14    241152\n",
      "   macro avg       0.46      0.44      0.14    241152\n",
      "weighted avg       0.82      0.14      0.17    241152\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_y, result))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.01588178519757344\n"
     ]
    }
   ],
   "source": [
    "print(cohen_kappa_score(test_y, result))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.09120499973004734\n"
     ]
    }
   ],
   "source": [
    "print(matthews_corrcoef(test_y, result))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}