natural-language-processing-spam-email-classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a href=\"https://www.facebook.com/codemakerz\"><img src=\"https://scontent.ffjr1-4.fna.fbcdn.net/v/t1.0-9/36189148_736466693143793_2172101683281133568_n.png?_nc_cat=107&_nc_eui2=AeHzxv3SUcQBOfijLP-cEnHkX4z9XQXdeau__2MlErWZ1x07aZ1zx1PzJUDDxL6cpr7oPqYiifggXDptgtP8W5iCoDRjcdILDBYZ5Ig40dqi8Q&_nc_oc=AQmMCNXdzelFB2rdtpk8wN8nC410Wm2yKupYfYS1FxHNejTF0Jhr1G3WIZORKRF3TvFpohMB8Puw29Txxan8CW05&_nc_ht=scontent.ffjr1-4.fna&oh=7b13627e991a4d1b508923041bd7bc22&oe=5D8A7B03\" />\n",
    "</a>\n",
    "Follow Us:\n",
    "Facebook: https://www.facebook.com/codemakerz\n",
    "\n",
    "<h1>Natural Language Processing - Email Ham & Spam</h1>\n",
    "<h3>As a learner i am also looking for new things, Help us with your suggestion and ideas. </h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/kaggle/input/spam-and-ham/spam.csv\n"
     ]
    }
   ],
   "source": [
    "# This Python 3 environment comes with many helpful analytics libraries installed\n",
    "# It is defined by the kaggle/python docker image: https://github.com/kaggle/docker-python\n",
    "# For example, here's several helpful packages to load in \n",
    "\n",
    "import numpy as np # linear algebra\n",
    "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
    "from nltk import FreqDist\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "# Input data files are available in the \"../input/\" directory.\n",
    "# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n",
    "\n",
    "import os\n",
    "for dirname, _, filenames in os.walk('/kaggle/input'):\n",
    "    for filename in filenames:\n",
    "        print(os.path.join(dirname, filename))\n",
    "\n",
    "# Any results you write to the current directory are saved as output."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a",
    "collapsed": true
   },
   "source": [
    "# Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_mails = pd.read_csv('/kaggle/input/spam-and-ham/spam.csv',encoding= 'latin-1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>v1</th>\n",
       "      <th>v2</th>\n",
       "      <th>Unnamed: 2</th>\n",
       "      <th>Unnamed: 3</th>\n",
       "      <th>Unnamed: 4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>ham</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>ham</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>spam</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>ham</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>ham</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     v1                                                 v2 Unnamed: 2  \\\n",
       "0   ham  Go until jurong point, crazy.. Available only ...        NaN   \n",
       "1   ham                      Ok lar... Joking wif u oni...        NaN   \n",
       "2  spam  Free entry in 2 a wkly comp to win FA Cup fina...        NaN   \n",
       "3   ham  U dun say so early hor... U c already then say...        NaN   \n",
       "4   ham  Nah I don't think he goes to usf, he lives aro...        NaN   \n",
       "\n",
       "  Unnamed: 3 Unnamed: 4  \n",
       "0        NaN        NaN  \n",
       "1        NaN        NaN  \n",
       "2        NaN        NaN  \n",
       "3        NaN        NaN  \n",
       "4        NaN        NaN  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_mails.head() # HAM - GOOD EMAILS, SPAM - BAD EMAILS"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Missing Values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "v1               0\n",
       "v2               0\n",
       "Unnamed: 2    5522\n",
       "Unnamed: 3    5560\n",
       "Unnamed: 4    5566\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_mails.isnull().sum()  #There are no missing value except unnamed column, we dont need those cols."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ham     4825\n",
       "spam     747\n",
       "Name: v1, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_mails.v1.value_counts()  # We can see there are total 747 spam and 4825 ham"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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": [
    "df_mails.v1.value_counts().plot(kind=\"bar\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>v1</th>\n",
       "      <th>v2</th>\n",
       "      <th>Unnamed: 2</th>\n",
       "      <th>Unnamed: 3</th>\n",
       "      <th>Unnamed: 4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>spam</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>spam</td>\n",
       "      <td>FreeMsg Hey there darling it's been 3 week's n...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>spam</td>\n",
       "      <td>WINNER!! As a valued network customer you have...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>spam</td>\n",
       "      <td>Had your mobile 11 months or more? U R entitle...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>spam</td>\n",
       "      <td>SIX chances to win CASH! From 100 to 20,000 po...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      v1                                                 v2 Unnamed: 2  \\\n",
       "2   spam  Free entry in 2 a wkly comp to win FA Cup fina...        NaN   \n",
       "5   spam  FreeMsg Hey there darling it's been 3 week's n...        NaN   \n",
       "8   spam  WINNER!! As a valued network customer you have...        NaN   \n",
       "9   spam  Had your mobile 11 months or more? U R entitle...        NaN   \n",
       "11  spam  SIX chances to win CASH! From 100 to 20,000 po...        NaN   \n",
       "\n",
       "   Unnamed: 3 Unnamed: 4  \n",
       "2         NaN        NaN  \n",
       "5         NaN        NaN  \n",
       "8         NaN        NaN  \n",
       "9         NaN        NaN  \n",
       "11        NaN        NaN  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Find most popular spam token\n",
    "df_spam = df_mails[df_mails.v1 == 'spam']\n",
    "df_spam.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import Spacy for tokenization\n",
    "import spacy\n",
    "nlp = spacy.load(\"en_core_web_sm\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\".\".isalpha()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "spam_token=[]\n",
    "famous_keyword = []\n",
    "for spam in np.array(df_spam.v2):\n",
    "    doc = nlp(spam.lower())\n",
    "    for token in doc:\n",
    "        # add famous keywords\n",
    "        if token.pos_ == \"NOUN\" or token.pos_ == 'PRON' or token.pos_ == 'PROPN':\n",
    "            if not token.text in famous_keyword and not token.is_stop and token.text.isalpha():\n",
    "                famous_keyword.append(token.text)\n",
    "        # add all spam tokens                              \n",
    "        if not token.is_stop and not token.text.isdigit() and token.text.isalpha():\n",
    "            spam_token.append(token.text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['entry', 'comp', 'cup', 'tkts', 'text', 'fa', 'txt', 'apply', 'week', 'word']"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# So these are keywords which u will get usually in spam messages\n",
    "famous_keyword[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['free', 'entry', 'wkly', 'comp', 'win', 'fa', 'cup', 'final', 'tkts', 'text']"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "spam_token[0:10] # so these are most common unique spam keywords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "FreqDist({'free': 222, 'txt': 154, 'ur': 144, 'u': 137, 'mobile': 127, 'text': 123, 'stop': 116, 'claim': 113, 'reply': 101, 'prize': 92, ...})"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Frequency Distribution\n",
    "freq_spam = FreqDist(spam_token)\n",
    "freq_spam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# so we can see that mostly spam keywrods are entry, free, prize, claim which make sense. In regular life when\n",
    "# you recieve any spam message they include these keywords.\n",
    "plt.figure(figsize=(15,10))\n",
    "freq_spam.plot(50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Split Train & Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "import re # for regula rexpression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# before splitting data we will try to reduce the dimensionality of tfid matrix by filtering stop words and \n",
    "# lemmatization.\n",
    "\n",
    "corpus = []\n",
    "for i in range(df_mails.shape[0]):\n",
    "    msg = re.sub('[^a-zA-Z]', ' ', df_mails.v2[i] ) # we will remove all the special characters\n",
    "    msg = msg.lower() # change it to lower case\n",
    "    doc = nlp(msg) # create spacy document\n",
    "    # remove stop words and perform. lemmatization\n",
    "    tokens_no_stop = [token.lemma_ for token in doc if not token.is_stop and not token.text.isspace()]\n",
    "    msg = ' '.join(tokens_no_stop) # join all the tokens to make sentence.\n",
    "    corpus.append(msg) # append to corpus list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['jurong point crazy available bugi n great world la e buffet cine get amore wat',\n",
       " 'ok lar joke wif u oni',\n",
       " 'free entry wkly comp win fa cup final tkts st text fa receive entry question std txt rate t c s apply s',\n",
       " 'u dun early hor u c',\n",
       " 'nah don t think go usf live',\n",
       " 'freemsg hey darle s week s word d like fun tb ok xxx std chgs send rcv',\n",
       " 'brother like speak treat like aid patent',\n",
       " 'request melle melle oru minnaminunginte nurungu vettam set callertune caller press copy friend callertune',\n",
       " 'winner value network customer select receivea prize reward claim claim code kl valid hour',\n",
       " 'mobile month u r entitle update late colour mobile camera free mobile update co free']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Above whole step is to reduce the dimensionality nd to provide only valid text to you model.\n",
    "corpus[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = corpus # Email\n",
    "y = df_mails[\"v1\"] # Result Ham or Spam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.33, random_state=34)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Term Frequency Inverse Document"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorizer = TfidfVectorizer()\n",
    "X_train_tfidf = vectorizer.fit_transform(X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "          intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
       "          multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
       "          verbose=0)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.svm import LinearSVC\n",
    "clf = LinearSVC()\n",
    "clf.fit(X_train_tfidf,y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pipeline"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "We applied vecotirzer for train dataset, same we have to apply for test dataset also, In our case there are less step. But there may be case where you have to apply tokenization, lemmatization, stemming, stop words. In this case you have apply all the step again which will take an extra effort. So better way to do it is defining a pipeline. So next time when we need to repeat those step, we can simply call that pipeline."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('tfidf',\n",
       "                 TfidfVectorizer(analyzer='word', binary=False,\n",
       "                                 decode_error='strict',\n",
       "                                 dtype=<class 'numpy.float64'>,\n",
       "                                 encoding='utf-8', input='content',\n",
       "                                 lowercase=True, max_df=1.0, max_features=None,\n",
       "                                 min_df=1, ngram_range=(1, 1), norm='l2',\n",
       "                                 preprocessor=None, smooth_idf=True,\n",
       "                                 stop_words=None, strip_accents=None,\n",
       "                                 sublinear_tf=False,\n",
       "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
       "                                 tokenizer=None, use_idf=True,\n",
       "                                 vocabulary=None)),\n",
       "                ('clf',\n",
       "                 LinearSVC(C=1.0, class_weight=None, dual=True,\n",
       "                           fit_intercept=True, intercept_scaling=1,\n",
       "                           loss='squared_hinge', max_iter=1000,\n",
       "                           multi_class='ovr', penalty='l2', random_state=None,\n",
       "                           tol=0.0001, verbose=0))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text_clf = Pipeline([('tfidf', TfidfVectorizer()),\n",
    "                     ('clf', LinearSVC()),\n",
    "])\n",
    "\n",
    "# Feed the training data through the pipeline\n",
    "text_clf.fit(X_train, y_train)  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = text_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['ham', 'ham', 'ham', ..., 'ham', 'ham', 'ham'], dtype=object)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions # See prediction is in text form whichis very good. IN other machine learning algorithm we\n",
    "# need to apply labelencoder."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Confusion Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7ff8d16fb438>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "cm = metrics.confusion_matrix(predictions, y_test)\n",
    "\n",
    "plt.figure(figsize=(10, 8))\n",
    "sns.heatmap(cm, annot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classification Report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         ham       0.98      0.99      0.99      1594\n",
      "        spam       0.96      0.89      0.93       245\n",
      "\n",
      "    accuracy                           0.98      1839\n",
      "   macro avg       0.97      0.94      0.96      1839\n",
      "weighted avg       0.98      0.98      0.98      1839\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test,predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# So we can see we are getting a good prediction and recall for both the cases(ham & spam)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9809679173463839\n"
     ]
    }
   ],
   "source": [
    "print(metrics.accuracy_score(y_test, predictions)) # we got an accuray of 98% which is really amazing."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test New Email"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['spam'], dtype=object)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text_clf.predict([\"Weekly Lottery Participation. Win upto $10,000.\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['ham'], dtype=object)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text_clf.predict([\"Hello Sir. How are you?\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# So it is working as we expected. Try some more messages. !!! \n",
    "# Thank you... !! UPVOTE if you like the code. "
   ]
  }
 ],
 "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}