Skin Cancer Detection 1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "finite-exhibition",
   "metadata": {},
   "source": [
    "## Importing data & libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "arabic-encyclopedia",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import cv2\n",
    "import json\n",
    "import random\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from glob import glob\n",
    "from PIL import Image\n",
    "from numpy.random import seed\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import imshow\n",
    "\n",
    "from keras.models import Model\n",
    "from keras import layers,models\n",
    "from keras.models import Sequential\n",
    "from keras.preprocessing import image\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.layers import Input, Lambda, Dense, Flatten\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "from keras.applications.inception_v3 import InceptionV3, preprocess_input\n",
    "from keras.preprocessing.image import array_to_img, img_to_array, load_img"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "homeless-agenda",
   "metadata": {},
   "source": [
    "## Preparing the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "average-western",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('./input/skin-metadata.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "emotional-somewhere",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10015, 7)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "african-yukon",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "nv       6705\n",
       "mel      1113\n",
       "bkl      1099\n",
       "bcc       514\n",
       "akiec     327\n",
       "vasc      142\n",
       "df        115\n",
       "Name: dx, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['dx'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "adapted-astrology",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='dx', ylabel='count'>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEJCAYAAABlmAtYAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAXoUlEQVR4nO3df7RddXnn8ffHoIAOqTAEigk0tJOqgKM01xSHaYuiEvtDGBWNM5RMpROHoahr1VEYZ1qnTlqmtbNGHGGGUiVRW0yxSnQExbToVNF4g0gIP0oUhJSUxF8D1rXQxGf+ON+Mx5uTu2/wnnNuzPu11lln7+fs7znPubk3n7v3d599U1VIkjSdJ4y7AUnS3GdYSJI6GRaSpE6GhSSpk2EhSepkWEiSOg0tLJI8PcltfbdHkrwhyVFJbkpyb7s/sm/MpUm2JrknyVl99aVJNrfHLk+SYfUtSdpbRvE5iyTzgL8Dfh64CPhGVV2W5BLgyKp6c5KTgD8HlgFPAz4J/GxV7U6yEXg98DngY8DlVXXD0BuXJAFwyIhe50zgy1X11SRnA2e0+hrgZuDNwNnAtVX1GHBfkq3AsiT3A/Or6haAJGuBc4Bpw+Loo4+uxYsXz/obkaQfZ5s2bfpaVS2YWh9VWKygt9cAcGxVbQeoqu1Jjmn1hfT2HPbY1mrfa8tT69NavHgxk5OTP2rfknRQSfLVQfWhT3AneRLwUuAvujYdUKtp6oNea1WSySSTO3fu3L9GJUn7NIqzoV4C3FpVD7f1h5McB9Dud7T6NuD4vnGLgIdafdGA+l6q6qqqmqiqiQUL9tqLkiQ9TqMIi1fzg0NQAOuBlW15JXB9X31FkkOTnAgsATa2Q1aPJjmtnQV1ft8YSdIIDHXOIsmTgRcBr+0rXwasS3IB8ABwLkBVbUmyDrgT2AVcVFW725gLgWuAw+lNbHsmlCSN0EhOnR2HiYmJcoJbkvZPkk1VNTG17ie4JUmdDAtJUifDQpLUybCQJHUa1Se49SN64PeeNe4W9umE39k87hYkDZl7FpKkToaFJKmTYSFJ6mRYSJI6GRaSpE6GhSSpk2EhSepkWEiSOhkWkqROhoUkqZNhIUnqZFhIkjoZFpKkToaFJKmTYSFJ6mRYSJI6GRaSpE5DDYskT01yXZK7k9yV5HlJjkpyU5J72/2RfdtfmmRrknuSnNVXX5pkc3vs8iQZZt+SpB827D2LdwA3VtUzgGcDdwGXABuqagmwoa2T5CRgBXAysBy4Ism89jxXAquAJe22fMh9S5L6DC0skswHfhH4U4Cq+m5VfQs4G1jTNlsDnNOWzwaurarHquo+YCuwLMlxwPyquqWqCljbN0aSNALD3LP4aWAn8J4kX0xydZKnAMdW1XaAdn9M234h8GDf+G2ttrAtT63vJcmqJJNJJnfu3Dm770aSDmLDDItDgJ8DrqyqU4F/oB1y2odB8xA1TX3vYtVVVTVRVRMLFizY334lSfswzLDYBmyrqs+39evohcfD7dAS7X5H3/bH941fBDzU6osG1CVJIzK0sKiqvwceTPL0VjoTuBNYD6xstZXA9W15PbAiyaFJTqQ3kb2xHap6NMlp7Syo8/vGSJJG4JAhP//FwPuTPAn4CvAb9AJqXZILgAeAcwGqakuSdfQCZRdwUVXtbs9zIXANcDhwQ7tJkkZkqGFRVbcBEwMeOnMf268GVg+oTwKnzGpzkqQZ8xPckqROhoUkqZNhIUnqZFhIkjoZFpKkToaFJKmTYSFJ6mRYSJI6GRaSpE6GhSSpk2EhSepkWEiSOhkWkqROhoUkqZNhIUnqZFhIkjoZFpKkToaFJKmTYSFJ6mRYSJI6GRaSpE6GhSSp01DDIsn9STYnuS3JZKsdleSmJPe2+yP7tr80ydYk9yQ5q6++tD3P1iSXJ8kw+5Yk/bBR7Fk8v6qeU1UTbf0SYENVLQE2tHWSnASsAE4GlgNXJJnXxlwJrAKWtNvyEfQtSWrGcRjqbGBNW14DnNNXv7aqHquq+4CtwLIkxwHzq+qWqipgbd8YSdIIDDssCvhEkk1JVrXasVW1HaDdH9PqC4EH+8Zua7WFbXlqfS9JViWZTDK5c+fOWXwbknRwO2TIz396VT2U5BjgpiR3T7PtoHmImqa+d7HqKuAqgImJiYHbSJL231D3LKrqoXa/A/gQsAx4uB1aot3vaJtvA47vG74IeKjVFw2oS5JGZGhhkeQpSY7Yswy8GLgDWA+sbJutBK5vy+uBFUkOTXIivYnsje1Q1aNJTmtnQZ3fN0aSNALDPAx1LPChdpbrIcCfVdWNSb4ArEtyAfAAcC5AVW1Jsg64E9gFXFRVu9tzXQhcAxwO3NBukqQRGVpYVNVXgGcPqH8dOHMfY1YDqwfUJ4FTZrtHSdLM+AluSVInw0KS1MmwkCR1MiwkSZ0MC0lSJ8NCktTJsJAkdTIsJEmdDAtJUifDQpLUybCQJHUyLCRJnQwLSVInw0KS1MmwkCR1MiwkSZ0MC0lSJ8NCktTJsJAkdTIsJEmdDAtJUifDQpLUaehhkWReki8m+WhbPyrJTUnubfdH9m17aZKtSe5JclZffWmSze2xy5Nk2H1Lkn5gFHsWrwfu6lu/BNhQVUuADW2dJCcBK4CTgeXAFUnmtTFXAquAJe22fAR9S5KaoYZFkkXArwBX95XPBta05TXAOX31a6vqsaq6D9gKLEtyHDC/qm6pqgLW9o2RJI3AsPcs/jvwJuD7fbVjq2o7QLs/ptUXAg/2bbet1Ra25an1vSRZlWQyyeTOnTtn5Q1IkoYYFkl+FdhRVZtmOmRAraap712suqqqJqpqYsGCBTN8WUlSl0OG+NynAy9N8svAYcD8JO8DHk5yXFVtb4eYdrTttwHH941fBDzU6osG1CVJIzK0PYuqurSqFlXVYnoT139VVecB64GVbbOVwPVteT2wIsmhSU6kN5G9sR2qejTJae0sqPP7xkiSRmBGYZFkw0xqM3QZ8KIk9wIvautU1RZgHXAncCNwUVXtbmMupDdJvhX4MnDD43xtSdLjMO1hqCSHAU8Gjm6fh9gzfzAfeNpMX6SqbgZubstfB87cx3argdUD6pPAKTN9PUnS7Oqas3gt8AZ6wbCJH4TFI8C7hteWJGkumTYsquodwDuSXFxV7xxRT5KkOWZGZ0NV1TuT/DNgcf+Yqlo7pL4kSXPIjMIiyXuBnwFuA/ZMOu/5NLUk6cfcTD9nMQGc1C63IUk6yMz0cxZ3AD85zEYkSXPXTPcsjgbuTLIReGxPsapeOpSuJElzykzD4q3DbEKSNLfN9GyoTw27EUnS3DXTs6Ee5QdXen0S8ETgH6pq/rAakyTNHTPdsziifz3JOcCyYTQkSZp7HtdVZ6vqw8ALZrcVSdJcNdPDUC/rW30Cvc9d+JkLSTpIzPRsqF/rW94F3E/vb2ZLkg4CM52z+I1hNyJJmrtm+sePFiX5UJIdSR5O8sEki7pHSpJ+HMx0gvs99P7s6dOAhcBHWk2SdBCYaVgsqKr3VNWudrsGWDDEviRJc8hMw+JrSc5LMq/dzgO+PszGJElzx0zD4jXAK4G/B7YDrwCc9Jakg8RMT519G7Cyqr4JkOQo4O30QkSS9GNupnsW/3RPUABU1TeAU4fTkiRprplpWDwhyZF7VtqexbR7JUkOS7IxyZeSbEnyn/eMTXJTknvbff/zXppka5J7kpzVV1+aZHN77PIk2b+3KUn6Ucw0LP4Y+GyStyX5PeCzwB92jHkMeEFVPRt4DrA8yWnAJcCGqloCbGjrJDkJWAGcDCwHrkgyrz3XlcAqYEm7LZ9h35KkWTCjsKiqtcDLgYeBncDLquq9HWOqqr7dVp/YbkXvMiFrWn0NcE5bPhu4tqoeq6r7gK3AsiTHAfOr6pb2N8DX9o2RJI3ATCe4qao7gTv358nbnsEm4J8A76qqzyc5tqq2t+fcnuSYtvlC4HN9w7e12vfa8tT6oNdbRW8PhBNOOGF/WpUkTeNxXaJ8pqpqd1U9B1hEby/hlGk2HzQPUdPUB73eVVU1UVUTCxb4mUFJmi1DDYs9qupbwM305hoeboeWaPc72mbbgOP7hi0CHmr1RQPqkqQRGVpYJFmQ5Klt+XDghcDd9K4xtbJtthK4vi2vB1YkOTTJifQmsje2Q1aPJjmtnQV1ft8YSdIIzHjO4nE4DljT5i2eAKyrqo8muQVYl+QC4AHgXICq2pJkHb15kV3ARVW1uz3XhcA1wOHADe0mSRqRoYVFVd3OgA/uVdXXgTP3MWY1sHpAfRKYbr5DkjREI5mzkCQd2AwLSVInw0KS1MmwkCR1MiwkSZ0MC0lSJ8NCktTJsJAkdTIsJEmdDAtJUifDQpLUybCQJHUyLCRJnQwLSVInw0KS1MmwkCR1MiwkSZ0MC0lSJ8NCktTJsJAkdTIsJEmdDAtJUqehhUWS45P8dZK7kmxJ8vpWPyrJTUnubfdH9o25NMnWJPckOauvvjTJ5vbY5UkyrL4lSXsb5p7FLuC3q+qZwGnARUlOAi4BNlTVEmBDW6c9tgI4GVgOXJFkXnuuK4FVwJJ2Wz7EviVJUwwtLKpqe1Xd2pYfBe4CFgJnA2vaZmuAc9ry2cC1VfVYVd0HbAWWJTkOmF9Vt1RVAWv7xkiSRmAkcxZJFgOnAp8Hjq2q7dALFOCYttlC4MG+YdtabWFbnlof9Dqrkkwmmdy5c+esvgdJOpgNPSyS/CPgg8AbquqR6TYdUKtp6nsXq66qqomqmliwYMH+NytJGmioYZHkifSC4v1V9Zet/HA7tES739Hq24Dj+4YvAh5q9UUD6pKkERnm2VAB/hS4q6r+W99D64GVbXklcH1ffUWSQ5OcSG8ie2M7VPVoktPac57fN0aSNAKHDPG5Twd+Hdic5LZW+w/AZcC6JBcADwDnAlTVliTrgDvpnUl1UVXtbuMuBK4BDgduaDdJ0ogMLSyq6m8YPN8AcOY+xqwGVg+oTwKnzF53kqT94Se4JUmdDAtJUifDQpLUybCQJHUyLCRJnQwLSVInw0KS1MmwkCR1MiwkSZ0MC0lSJ8NCktTJsJAkdTIsJEmdDAtJUifDQpLUybCQJHUyLCRJnQwLSVInw0KS1MmwkCR1MiwkSZ0OGXcD0oHgU7/4S+NuYaBf+vSnxt2CDhJDC4sk7wZ+FdhRVae02lHAB4DFwP3AK6vqm+2xS4ELgN3A66rq462+FLgGOBz4GPD6qqr97Wfpv1/7o72hIdr0R+ePuwVJmtYwD0NdAyyfUrsE2FBVS4ANbZ0kJwErgJPbmCuSzGtjrgRWAUvabepzSpKGbGhhUVWfBr4xpXw2sKYtrwHO6atfW1WPVdV9wFZgWZLjgPlVdUvbm1jbN0aSNCKjnuA+tqq2A7T7Y1p9IfBg33bbWm1hW55aHyjJqiSTSSZ37tw5q41L0sFsrpwNlQG1mqY+UFVdVVUTVTWxYMGCWWtOkg52ow6Lh9uhJdr9jlbfBhzft90i4KFWXzSgLkkaoVGHxXpgZVteCVzfV1+R5NAkJ9KbyN7YDlU9muS0JAHO7xsjSRqRYZ46++fAGcDRSbYBvwtcBqxLcgHwAHAuQFVtSbIOuBPYBVxUVbvbU13ID06dvaHdJEkjNLSwqKpX7+OhM/ex/Wpg9YD6JHDKLLYmSdpPc2WCW5I0hxkWkqROhoUkqZNhIUnqZFhIkjoZFpKkToaFJKmTYSFJ6mRYSJI6+WdVJc15q897xbhbGOgt77tu3C2MjHsWkqROhoUkqZNhIUnqZFhIkjoZFpKkTp4NJUlDdNfqvxp3C/v0zLe8YMbbumchSepkWEiSOhkWkqROhoUkqZMT3NJB4H/89kfG3cJAv/XHvzbuFjRDB0xYJFkOvAOYB1xdVZeNuSXth9Pfefq4WxjoMxd/ZtwtSAeEA+IwVJJ5wLuAlwAnAa9OctJ4u5Kkg8cBERbAMmBrVX2lqr4LXAucPeaeJOmgcaCExULgwb71ba0mSRqBVNW4e+iU5FzgrKr6zbb+68Cyqrp4ynargFVt9enAPUNs62jga0N8/mE6kHsH+x83+x+vYff/U1W1YGrxQJng3gYc37e+CHho6kZVdRVw1SgaSjJZVROjeK3ZdiD3DvY/bvY/XuPq/0A5DPUFYEmSE5M8CVgBrB9zT5J00Dgg9iyqaleS3wI+Tu/U2XdX1ZYxtyVJB40DIiwAqupjwMfG3UefkRzuGpIDuXew/3Gz//EaS/8HxAS3JGm8DpQ5C0nSGBkW00iyOMkdA+r3Jzl6QP3bo+ns4JbkrUnemOQZSW5L8sUkPzPuvmZDkjOSfHTcfRzI9vVzeyAZ9H9JkqcluW4c/YBhoQPbOcD1VXVqVX153M1Iw1RVD1XVK8b1+oZFt0OSrElye5Lrkjx5zwNJDk9yY5J/M84Gp9N+y7oryZ8k2ZLkE0memWTjlG1uH2efXZK8Jck9ST5J7wOXTwbeAPxmkr8ea3NTtK/n3UmuTnJHkvcneWGSzyS5N8myJE9J8u4kX2h7RmO9fE2S/5rk3/WtvzXJ7ybZkOTWJJv39Nh6/99JvtTe36ta/blJPtvqG5McMa73w4Cf20H9JZmX5O3t/d2e5OLup55dST6cZFP7+Vw15bGjk9yS5Ff695ha33/Uvn9uT/LavjFvau/nS0lm74KrVeVtHzdgMVDA6W393cAbgfvbY58Ezu/b/tvj7nkf72EX8Jy2vg44D7gN+OlWezPwH8fd6zTvYSmwmV5AzAe2tn+HtwJvHHd/03zNn0XvF7JN7Xsn9K5p9mHg94Hz2vZPBf4WeApwBvDRMfR8KvCpvvU7gROA+W396PZ1D/By4E/6tv0J4EnAV4Dnttp84JAxfv2n/ty+aVB/wIXAB/f0Chw1hn6PaveHA3cA/xj4NnAs8HngRX3v6462vGrPzyxwKDAJnEjvYqufBZ482+/HPYtuD1bVnutYvw/45235euA9VbV2PG3tl/uq6ra2vIneN9064JWt9irgA6Nva8Z+AfhQVX2nqh7hwPhA5n1Vtbmqvg9sATZU76d3M72v/4uBS5LcBtwMHEbvP+exqKovAse04+LPBr4JbAd+v+11fpLe9diOpfceXtj2Rn6hqv4vvb297VX1hfZ8j1TVrrG8mZ6pP7dnMbi/FwL/c0+vVfWNMfT6uiRfAj5H70oVS4AnAhuAN1XVTQPGvBg4v33/fJ5ewCyh937eU1Xfgdl9PwfM5yzGaOq5xXvWPwO8JMmftf8E5rLH+pZ30/sN5r3AXyT5S6Cq6t6xdDZzc/1rPFX/1/z7fevfp/dztxt4eVX90PXLkhw7mvYGug54BfCT9K7s/K+ABcDSqvpekvuBw6rqb5MsBX4Z+IMkn6C3tzSX/o2m9vIIvd/Ap8qAbUcmyRn0/oN/XlV9J8nN9H5x2EXvF7uzgE8NGgpcXFUfn/J8yxnS+3HPotsJSZ7Xll8N/E1b/h3g68AVY+nqR1S9CeHdwH9ibu9VAHwa+BdtjugI4Mfhz6t9HLg4SQCSnDrmfqAXECvoBcZ19A4v7WhB8Xzgp6B3Vg7wnap6H/B24OeAu4GnJXlu2+aIJOP8ZXTqz+3nGNzfJ4B/u6fXJEeNuM+fAL7ZguIZwGmtXsBrgGckuWTAuI8DFyZ5IkCSn03yFHrv5zV75lZn8/0YFt3uAla2XfGjgCv7HnsDcFiSPxxHY7PgA/TmL9aNu5HpVNWt9Hq9jd7x5f8z1oZmx9voHWq4vU1avm3M/VC9S+gcAfxdVW0H3g9MJJmkt5dxd9v0WcDGdgjkLcB/qd7fmXkV8M52SOUmer8hj8vUn9t37qO/q4EH6P07fAn4lyPu80Z6k/G30/se+NyeB6pqN73wfn7/yQfN1fTmlW5t3z//i968y430DtNOtn+fN85Wo36CW5LUyT0LSVInw0KS1MmwkCR1MiwkSZ0MC0lSJ8NCGpF2vaVZO5VRGiXDQpLUybCQhih7Xy2XdqXQM9ryHyRZPcYWpRnx2lDSkLTrJ62gd0XXQ4Bb6V3v518D1yV5HbAc+Plx9SjNlGEhDc//v1ouQJL10LusRpL3Ah+hdwG5746xR2lGPAwlDde+rqfzLOBb9C75Lc15hoU0PAOvlpvkZfT+/sAvApcneer4WpRmxgsJSkOU5C3A+cBXgW30rhS6Cjizqh5s8xZLq2rlGNuUOhkWkqROHoaSJHUyLCRJnQwLSVInw0KS1MmwkCR1MiwkSZ0MC0lSJ8NCktTp/wHiWz2sHFtOJgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(df['dx'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "bored-soccer",
   "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>lesion_id</th>\n",
       "      <th>image_id</th>\n",
       "      <th>dx</th>\n",
       "      <th>dx_type</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>localization</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HAM_0000118</td>\n",
       "      <td>ISIC_0027419.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HAM_0000118</td>\n",
       "      <td>ISIC_0025030.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HAM_0002730</td>\n",
       "      <td>ISIC_0026769.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HAM_0002730</td>\n",
       "      <td>ISIC_0025661.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HAM_0001466</td>\n",
       "      <td>ISIC_0031633.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>75.0</td>\n",
       "      <td>male</td>\n",
       "      <td>ear</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     lesion_id          image_id   dx dx_type   age   sex localization\n",
       "0  HAM_0000118  ISIC_0027419.jpg  bkl   histo  80.0  male        scalp\n",
       "1  HAM_0000118  ISIC_0025030.jpg  bkl   histo  80.0  male        scalp\n",
       "2  HAM_0002730  ISIC_0026769.jpg  bkl   histo  80.0  male        scalp\n",
       "3  HAM_0002730  ISIC_0025661.jpg  bkl   histo  80.0  male        scalp\n",
       "4  HAM_0001466  ISIC_0031633.jpg  bkl   histo  75.0  male          ear"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "brief-austin",
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = df[['image_id', 'dx']].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "herbal-explanation",
   "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>image_id</th>\n",
       "      <th>dx</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ISIC_0027419.jpg</td>\n",
       "      <td>bkl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ISIC_0025030.jpg</td>\n",
       "      <td>bkl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ISIC_0026769.jpg</td>\n",
       "      <td>bkl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ISIC_0025661.jpg</td>\n",
       "      <td>bkl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ISIC_0031633.jpg</td>\n",
       "      <td>bkl</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           image_id   dx\n",
       "0  ISIC_0027419.jpg  bkl\n",
       "1  ISIC_0025030.jpg  bkl\n",
       "2  ISIC_0026769.jpg  bkl\n",
       "3  ISIC_0025661.jpg  bkl\n",
       "4  ISIC_0031633.jpg  bkl"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "welcome-mathematics",
   "metadata": {},
   "source": [
    "## Images to Array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "graphic-giving",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['HAM10000_metadata.numbers', 'HAM10000-metadata.csv', '.DS_Store', 'images_dataset', 'HAM10000_images_part_1', 'HAM10000_metadata2.csv', 'HAM_metadata.csv', 'HAM10000_metadata2.numbers', 'skin-metadata.csv', 'skin-metadata copy.csv', '.ipynb_checkpoints', 'HAM10000_images_part_2', 'HAM10000_metadata.csv']\n"
     ]
    }
   ],
   "source": [
    "print(os.listdir('./input/'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "vocal-grade",
   "metadata": {},
   "outputs": [],
   "source": [
    "path = './input/'\n",
    "images_path = {os.path.splitext(os.path.basename(x))[0]: x for x in glob(os.path.join(path, '*', '*.jpg'))}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "typical-train",
   "metadata": {},
   "outputs": [],
   "source": [
    "lesion_type_dict = {\n",
    "    'nv': 'Melanocytic nevi',\n",
    "    'mel': 'Melanoma',\n",
    "    'bkl': 'Benign keratosis-like lesions ',\n",
    "    'bcc': 'Basal cell carcinoma',\n",
    "    'akiec': 'Actinic keratoses',\n",
    "    'vasc': 'Vascular lesions',\n",
    "    'df': 'Dermatofibroma'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "excess-wrong",
   "metadata": {},
   "outputs": [],
   "source": [
    "skin_df = pd.read_csv(os.path.join(path, 'HAM10000-metadata.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "placed-worcester",
   "metadata": {},
   "outputs": [],
   "source": [
    "skin_df['path'] = skin_df['image_id'].map(images_path.get)\n",
    "skin_df['cell_type'] = skin_df['dx'].map(lesion_type_dict.get)\n",
    "skin_df['cell_type_idx'] = pd.Categorical(skin_df['cell_type']).codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "acquired-catalog",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 10015 entries, 0 to 10014\n",
      "Data columns (total 10 columns):\n",
      " #   Column         Non-Null Count  Dtype  \n",
      "---  ------         --------------  -----  \n",
      " 0   lesion_id      10015 non-null  object \n",
      " 1   image_id       10015 non-null  object \n",
      " 2   dx             10015 non-null  object \n",
      " 3   dx_type        10015 non-null  object \n",
      " 4   age            9958 non-null   float64\n",
      " 5   sex            10015 non-null  object \n",
      " 6   localization   10015 non-null  object \n",
      " 7   path           10015 non-null  object \n",
      " 8   cell_type      10015 non-null  object \n",
      " 9   cell_type_idx  10015 non-null  int8   \n",
      "dtypes: float64(1), int8(1), object(8)\n",
      "memory usage: 714.1+ KB\n"
     ]
    }
   ],
   "source": [
    "skin_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "private-tooth",
   "metadata": {},
   "outputs": [],
   "source": [
    "skin_df['image'] = skin_df['path'].map(lambda x: np.asarray(Image.open(x).resize((75,75))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "public-compensation",
   "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>lesion_id</th>\n",
       "      <th>image_id</th>\n",
       "      <th>dx</th>\n",
       "      <th>dx_type</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>localization</th>\n",
       "      <th>path</th>\n",
       "      <th>cell_type</th>\n",
       "      <th>cell_type_idx</th>\n",
       "      <th>image</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HAM_0000118</td>\n",
       "      <td>ISIC_0027419</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0027419.jpg</td>\n",
       "      <td>Benign keratosis-like lesions</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[191, 153, 195], [191, 154, 195], [192, 150,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HAM_0000118</td>\n",
       "      <td>ISIC_0025030</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0025030.jpg</td>\n",
       "      <td>Benign keratosis-like lesions</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[24, 13, 22], [24, 14, 26], [30, 18, 37], [4...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HAM_0002730</td>\n",
       "      <td>ISIC_0026769</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0026769.jpg</td>\n",
       "      <td>Benign keratosis-like lesions</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[186, 128, 139], [191, 134, 149], [197, 143,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HAM_0002730</td>\n",
       "      <td>ISIC_0025661</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0025661.jpg</td>\n",
       "      <td>Benign keratosis-like lesions</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[23, 11, 18], [31, 16, 26], [52, 30, 40], [7...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HAM_0001466</td>\n",
       "      <td>ISIC_0031633</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>75.0</td>\n",
       "      <td>male</td>\n",
       "      <td>ear</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0031633.jpg</td>\n",
       "      <td>Benign keratosis-like lesions</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[136, 93, 115], [153, 108, 132], [170, 126, ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     lesion_id      image_id   dx dx_type   age   sex localization  \\\n",
       "0  HAM_0000118  ISIC_0027419  bkl   histo  80.0  male        scalp   \n",
       "1  HAM_0000118  ISIC_0025030  bkl   histo  80.0  male        scalp   \n",
       "2  HAM_0002730  ISIC_0026769  bkl   histo  80.0  male        scalp   \n",
       "3  HAM_0002730  ISIC_0025661  bkl   histo  80.0  male        scalp   \n",
       "4  HAM_0001466  ISIC_0031633  bkl   histo  75.0  male          ear   \n",
       "\n",
       "                                              path  \\\n",
       "0  ./input/HAM10000_images_part_1/ISIC_0027419.jpg   \n",
       "1  ./input/HAM10000_images_part_1/ISIC_0025030.jpg   \n",
       "2  ./input/HAM10000_images_part_1/ISIC_0026769.jpg   \n",
       "3  ./input/HAM10000_images_part_1/ISIC_0025661.jpg   \n",
       "4  ./input/HAM10000_images_part_1/ISIC_0031633.jpg   \n",
       "\n",
       "                        cell_type  cell_type_idx  \\\n",
       "0  Benign keratosis-like lesions               2   \n",
       "1  Benign keratosis-like lesions               2   \n",
       "2  Benign keratosis-like lesions               2   \n",
       "3  Benign keratosis-like lesions               2   \n",
       "4  Benign keratosis-like lesions               2   \n",
       "\n",
       "                                               image  \n",
       "0  [[[191, 153, 195], [191, 154, 195], [192, 150,...  \n",
       "1  [[[24, 13, 22], [24, 14, 26], [30, 18, 37], [4...  \n",
       "2  [[[186, 128, 139], [191, 134, 149], [197, 143,...  \n",
       "3  [[[23, 11, 18], [31, 16, 26], [52, 30, 40], [7...  \n",
       "4  [[[136, 93, 115], [153, 108, 132], [170, 126, ...  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "skin_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "accessible-hebrew",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "lesion_id         0\n",
       "image_id          0\n",
       "dx                0\n",
       "dx_type           0\n",
       "age              57\n",
       "sex               0\n",
       "localization      0\n",
       "path              0\n",
       "cell_type         0\n",
       "cell_type_idx     0\n",
       "image             0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "skin_df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "bacterial-manner",
   "metadata": {},
   "outputs": [],
   "source": [
    "skin_df = skin_df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "derived-pepper",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9958, 11)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "skin_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "noble-stadium",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.utils import shuffle\n",
    "skin_df = shuffle(skin_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "acute-teach",
   "metadata": {},
   "outputs": [],
   "source": [
    "skin_df.reset_index(inplace=True, drop=True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "plastic-province",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(75, 75, 3)    9958\n",
       "Name: image, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking the image size distribution\n",
    "\n",
    "skin_df['image'].map(lambda x: x.shape).value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "healthy-upgrade",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = skin_df['image']\n",
    "y = skin_df['cell_type_idx']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "human-republican",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9958,)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data normalization\n",
    "\n",
    "X = X.values\n",
    "X = X/255\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "revised-patio",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(9958, 75, 75, 3)\n"
     ]
    }
   ],
   "source": [
    "lst = []\n",
    "for _ in X:\n",
    "    lst.append(_)\n",
    "X = np.array(lst)\n",
    "print(X.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "obvious-manual",
   "metadata": {},
   "source": [
    "## Splitting the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "dutch-visit",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.20,random_state=28)\n",
    "x_train, x_val, y_train, y_val = train_test_split(x_train, y_train, test_size=0.20,random_state=28)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "abroad-delta",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils.np_utils import to_categorical\n",
    "\n",
    "y_test = to_categorical(y_test, num_classes = 7)\n",
    "y_train = to_categorical(y_train, num_classes = 7)\n",
    "y_val = to_categorical(y_val , num_classes=7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "conventional-notification",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X shape: (9958, 75, 75, 3)\n",
      "number of test examples = 1992\n",
      "Y_test shape: (1992, 7)\n"
     ]
    }
   ],
   "source": [
    "# We only require test data\n",
    "\n",
    "print (\"X shape: \" + str(X.shape))\n",
    "print (\"number of test examples = \" + str(y_test.shape[0]))\n",
    "print (\"Y_test shape: \" + str(y_test.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "later-nylon",
   "metadata": {},
   "source": [
    "## Data augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "id": "automotive-vintage",
   "metadata": {},
   "outputs": [],
   "source": [
    "datagen = ImageDataGenerator(\n",
    "        preprocessing_function=preprocess_input)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "id": "buried-photographer",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 10015 validated image filenames belonging to 7 classes.\n"
     ]
    }
   ],
   "source": [
    "train_generator = datagen.flow_from_dataframe(\n",
    "    df1,\n",
    "    directory = './input/HAM10000_images_part_1/',\n",
    "    x_col = 'image_id',\n",
    "    y_col = 'dx',\n",
    "    target_size = (75, 75),\n",
    "    class_mode = 'categorical',\n",
    "    batch_size = 32)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "wanted-vegetarian",
   "metadata": {},
   "source": [
    "## Building the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "id": "collective-rebate",
   "metadata": {},
   "outputs": [],
   "source": [
    "skinModel = InceptionV3(include_top = False, weights = 'imagenet', input_shape = (75, 75, 3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "id": "severe-viking",
   "metadata": {},
   "outputs": [],
   "source": [
    "skinModel.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "id": "tracked-purchase",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = models.Sequential()\n",
    "model.add(skinModel)\n",
    "\n",
    "model.add(layers.GlobalAveragePooling2D())\n",
    "model.add(layers.Dense(512, activation='relu'))\n",
    "\n",
    "model.add(layers.Dropout(0.5))\n",
    "\n",
    "model.add(layers.Dense(256, activation='relu'))\n",
    "model.add(layers.Dense(7, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "id": "immune-favorite",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "inception_v3 (Model)         (None, 1, 1, 2048)        21802784  \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_1 ( (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 512)               1049088   \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 256)               131328    \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 7)                 1799      \n",
      "=================================================================\n",
      "Total params: 22,984,999\n",
      "Trainable params: 1,182,215\n",
      "Non-trainable params: 21,802,784\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "id": "marine-curtis",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(\n",
    "    optimizer='adam',\n",
    "    loss = 'categorical_crossentropy',\n",
    "    metrics = ['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "id": "minor-constitutional",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_steps = np.ceil(train_generator.n / train_generator.batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "id": "literary-example",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/40\n",
      "313/313 [==============================] - 191s 609ms/step - loss: 0.8981 - accuracy: 0.6904\n",
      "Epoch 2/40\n",
      "313/313 [==============================] - 177s 567ms/step - loss: 0.8389 - accuracy: 0.7072\n",
      "Epoch 3/40\n",
      "313/313 [==============================] - 185s 592ms/step - loss: 0.7991 - accuracy: 0.7208\n",
      "Epoch 4/40\n",
      "313/313 [==============================] - 169s 541ms/step - loss: 0.7621 - accuracy: 0.7269\n",
      "Epoch 5/40\n",
      "313/313 [==============================] - 179s 572ms/step - loss: 0.7323 - accuracy: 0.7418\n",
      "Epoch 6/40\n",
      "313/313 [==============================] - 165s 526ms/step - loss: 0.6987 - accuracy: 0.7501\n",
      "Epoch 7/40\n",
      "313/313 [==============================] - 189s 603ms/step - loss: 0.6568 - accuracy: 0.7624\n",
      "Epoch 8/40\n",
      "313/313 [==============================] - 175s 559ms/step - loss: 0.6184 - accuracy: 0.7710\n",
      "Epoch 9/40\n",
      "313/313 [==============================] - 193s 615ms/step - loss: 0.5907 - accuracy: 0.7876\n",
      "Epoch 10/40\n",
      "313/313 [==============================] - 585s 2s/step - loss: 0.5696 - accuracy: 0.7925\n",
      "Epoch 11/40\n",
      "313/313 [==============================] - 189s 604ms/step - loss: 0.5416 - accuracy: 0.8013\n",
      "Epoch 12/40\n",
      "313/313 [==============================] - 190s 606ms/step - loss: 0.5098 - accuracy: 0.8165\n",
      "Epoch 13/40\n",
      "313/313 [==============================] - 174s 555ms/step - loss: 0.4886 - accuracy: 0.8212\n",
      "Epoch 14/40\n",
      "313/313 [==============================] - 190s 607ms/step - loss: 0.4515 - accuracy: 0.8321\n",
      "Epoch 15/40\n",
      "313/313 [==============================] - 190s 606ms/step - loss: 0.4310 - accuracy: 0.8442\n",
      "Epoch 16/40\n",
      "313/313 [==============================] - 204s 652ms/step - loss: 0.4123 - accuracy: 0.8491\n",
      "Epoch 17/40\n",
      "313/313 [==============================] - 193s 616ms/step - loss: 0.3944 - accuracy: 0.8528\n",
      "Epoch 18/40\n",
      "313/313 [==============================] - 192s 614ms/step - loss: 0.3660 - accuracy: 0.8619\n",
      "Epoch 19/40\n",
      "313/313 [==============================] - 189s 603ms/step - loss: 0.3489 - accuracy: 0.8701\n",
      "Epoch 20/40\n",
      "313/313 [==============================] - 184s 589ms/step - loss: 0.3491 - accuracy: 0.8715\n",
      "Epoch 21/40\n",
      "313/313 [==============================] - 185s 591ms/step - loss: 0.3312 - accuracy: 0.8805\n",
      "Epoch 22/40\n",
      "313/313 [==============================] - 186s 594ms/step - loss: 0.3133 - accuracy: 0.8877\n",
      "Epoch 23/40\n",
      "313/313 [==============================] - 184s 588ms/step - loss: 0.3010 - accuracy: 0.8928\n",
      "Epoch 24/40\n",
      "313/313 [==============================] - 190s 606ms/step - loss: 0.3041 - accuracy: 0.8905\n",
      "Epoch 25/40\n",
      "313/313 [==============================] - 200s 640ms/step - loss: 0.2868 - accuracy: 0.8970\n",
      "Epoch 26/40\n",
      "313/313 [==============================] - 186s 593ms/step - loss: 0.2737 - accuracy: 0.8997\n",
      "Epoch 27/40\n",
      "313/313 [==============================] - 205s 654ms/step - loss: 0.2598 - accuracy: 0.9072\n",
      "Epoch 28/40\n",
      "313/313 [==============================] - 201s 643ms/step - loss: 0.2676 - accuracy: 0.8986\n",
      "Epoch 29/40\n",
      "313/313 [==============================] - 197s 629ms/step - loss: 0.2514 - accuracy: 0.9117\n",
      "Epoch 30/40\n",
      "313/313 [==============================] - 199s 636ms/step - loss: 0.2421 - accuracy: 0.9124\n",
      "Epoch 31/40\n",
      "313/313 [==============================] - 194s 621ms/step - loss: 0.2284 - accuracy: 0.9163\n",
      "Epoch 32/40\n",
      "313/313 [==============================] - 191s 610ms/step - loss: 0.2411 - accuracy: 0.9138\n",
      "Epoch 33/40\n",
      "313/313 [==============================] - 198s 634ms/step - loss: 0.2299 - accuracy: 0.9168\n",
      "Epoch 34/40\n",
      "313/313 [==============================] - 188s 602ms/step - loss: 0.2079 - accuracy: 0.9262\n",
      "Epoch 35/40\n",
      "313/313 [==============================] - 187s 598ms/step - loss: 0.2109 - accuracy: 0.9204\n",
      "Epoch 36/40\n",
      "313/313 [==============================] - 192s 614ms/step - loss: 0.2070 - accuracy: 0.9246\n",
      "Epoch 37/40\n",
      "313/313 [==============================] - 194s 619ms/step - loss: 0.2092 - accuracy: 0.9253\n",
      "Epoch 38/40\n",
      "313/313 [==============================] - 179s 573ms/step - loss: 0.2101 - accuracy: 0.9228\n",
      "Epoch 39/40\n",
      "313/313 [==============================] - 181s 579ms/step - loss: 0.2062 - accuracy: 0.9251\n",
      "Epoch 40/40\n",
      "313/313 [==============================] - 177s 567ms/step - loss: 0.1826 - accuracy: 0.9347\n"
     ]
    }
   ],
   "source": [
    "from keras.callbacks import ModelCheckpoint, EarlyStopping\n",
    "\n",
    "monitor = EarlyStopping(monitor='val_loss', \n",
    "                        min_delta=1e-3, \n",
    "                        patience=5, verbose=1, \n",
    "                        restore_best_weights=True)\n",
    "\n",
    "history = model.fit(train_generator,\n",
    "                    epochs = 40,\n",
    "                    batch_size = 32,\n",
    "                    steps_per_epoch = train_steps)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cardiac-alexandria",
   "metadata": {},
   "source": [
    "## Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "sunrise-badge",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow import keras\n",
    "model1 = keras.models.load_model('./server/models/cancerModelV1.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "vulnerable-hygiene",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow import keras\n",
    "model2 = keras.models.load_model('./server/models/cancerModelV2.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "norwegian-debut",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63/63 [==============================] - 7s 112ms/step - loss: 9.1326 - accuracy: 0.1591\n",
      "\n",
      "Loss = 9.13256549835205\n",
      "Test Accuracy = 0.1591365486383438\n"
     ]
    }
   ],
   "source": [
    "preds = model1.evaluate(x_test, y_test, batch_size=32, verbose=1, sample_weight=None)\n",
    "print()\n",
    "print (\"Loss = \" + str(preds[0]))\n",
    "print (\"Test Accuracy = \" + str(preds[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "pharmaceutical-jumping",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63/63 [==============================] - 7s 113ms/step - loss: 10.7577 - accuracy: 0.1571\n",
      "\n",
      "Loss = 10.75768756866455\n",
      "Test Accuracy = 0.15712851285934448\n"
     ]
    }
   ],
   "source": [
    "preds = model2.evaluate(x_test, y_test, batch_size=32, verbose=1, sample_weight=None)\n",
    "print()\n",
    "print (\"Loss = \" + str(preds[0]))\n",
    "print (\"Test Accuracy = \" + str(preds[1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "front-nickname",
   "metadata": {},
   "source": [
    "## Test with single image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "grateful-waterproof",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[3.56966928e-02 1.09837905e-01 8.33620250e-01 1.16273435e-02\n",
      "  2.78383843e-03 6.43083779e-03 3.14920430e-06]]\n"
     ]
    },
    {
     "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": [
    "img_path = './input/HAM10000_images_part_1/ISIC_0032870.jpg'\n",
    "\n",
    "img = image.load_img(img_path, target_size=(75, 75))\n",
    "imshow(img)\n",
    "\n",
    "x = image.img_to_array(img)\n",
    "x = x/255\n",
    "x = x.reshape(1, 75, 75, 3)\n",
    "\n",
    "final_prediction = model2.predict(x)\n",
    "print(final_prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "designed-trout",
   "metadata": {},
   "outputs": [],
   "source": [
    "output_labels = [\"Actinic Keratoses and Intraepithelial Carcinoma / Bowen's Disease (akiec)\",\n",
    "                 \"Basal Cell Carcinoma (bcc)\",\n",
    "                 \"Benign Keratosis-like Lesions (solar lentigines / seborrheic keratoses and lichen-planus like keratoses, bkl)\",\n",
    "                 \"Dermatofibroma (df)\",\n",
    "                 \"Melanocytic Nevi (nv)\",\n",
    "                 \"Vascular Lesions (angiomas, angiokeratomas, pyogenic granulomas and hemorrhage, vasc)\",\n",
    "                 \"Melanoma (mel)\"]\n",
    "\n",
    "output_length = len(output_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "underlying-cheat",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Benign Keratosis-like Lesions (solar lentigines / seborrheic keratoses and lichen-planus like keratoses, bkl)\n"
     ]
    }
   ],
   "source": [
    "p = np.argmax(final_prediction[0])\n",
    "print(output_labels[p])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cultural-composite",
   "metadata": {},
   "source": [
    "## Model to feature extractor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "comic-biography",
   "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>lesion_id</th>\n",
       "      <th>image_id</th>\n",
       "      <th>dx</th>\n",
       "      <th>dx_type</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>localization</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HAM_0000118</td>\n",
       "      <td>ISIC_0027419.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HAM_0000118</td>\n",
       "      <td>ISIC_0025030.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HAM_0002730</td>\n",
       "      <td>ISIC_0026769.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HAM_0002730</td>\n",
       "      <td>ISIC_0025661.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>80.0</td>\n",
       "      <td>male</td>\n",
       "      <td>scalp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HAM_0001466</td>\n",
       "      <td>ISIC_0031633.jpg</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>75.0</td>\n",
       "      <td>male</td>\n",
       "      <td>ear</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     lesion_id          image_id   dx dx_type   age   sex localization\n",
       "0  HAM_0000118  ISIC_0027419.jpg  bkl   histo  80.0  male        scalp\n",
       "1  HAM_0000118  ISIC_0025030.jpg  bkl   histo  80.0  male        scalp\n",
       "2  HAM_0002730  ISIC_0026769.jpg  bkl   histo  80.0  male        scalp\n",
       "3  HAM_0002730  ISIC_0025661.jpg  bkl   histo  80.0  male        scalp\n",
       "4  HAM_0001466  ISIC_0031633.jpg  bkl   histo  75.0  male          ear"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "specialized-adrian",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "inception_v3_input (InputLay [(None, 75, 75, 3)]       0         \n",
      "_________________________________________________________________\n",
      "inception_v3 (Model)         (None, 1, 1, 2048)        21802784  \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_1 ( (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 512)               1049088   \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 512)               0         \n",
      "=================================================================\n",
      "Total params: 22,851,872\n",
      "Trainable params: 1,049,088\n",
      "Non-trainable params: 21,802,784\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model_a = Model(model2.input, model2.layers[3].output)\n",
    "model_a.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "romantic-bachelor",
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model_a.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "sweet-remainder",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.        , 0.        , 0.        , 3.06525   , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.1999509 , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       1.3238518 , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 3.1026974 , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 3.9476943 , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.45018068, 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 1.4244664 , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 1.056135  , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.5603839 ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 1.3272718 , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       3.5013075 , 0.        , 0.        , 0.        , 2.6977005 ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.40568325, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.32101136, 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 4.4304366 , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 1.304124  , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.44972917, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 2.69741   , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 3.7439537 ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.        , 0.        , 1.2268275 , 1.9293525 , 0.        ,\n",
       "       0.        , 0.        ], dtype=float32)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "governing-private",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "512"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(predictions[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "infrared-force",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9958, 512)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "ignored-multiple",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "numpy.ndarray"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "lightweight-phrase",
   "metadata": {},
   "outputs": [],
   "source": [
    "# New dataset that will include features\n",
    "\n",
    "data_df = skin_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "dressed-avenue",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Merge original dataframe with extracted features\n",
    "\n",
    "complete_data = pd.concat([data_df, pd.DataFrame(predictions)], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "graduate-permit",
   "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>lesion_id</th>\n",
       "      <th>image_id</th>\n",
       "      <th>dx</th>\n",
       "      <th>dx_type</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>localization</th>\n",
       "      <th>path</th>\n",
       "      <th>cell_type</th>\n",
       "      <th>cell_type_idx</th>\n",
       "      <th>...</th>\n",
       "      <th>502</th>\n",
       "      <th>503</th>\n",
       "      <th>504</th>\n",
       "      <th>505</th>\n",
       "      <th>506</th>\n",
       "      <th>507</th>\n",
       "      <th>508</th>\n",
       "      <th>509</th>\n",
       "      <th>510</th>\n",
       "      <th>511</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HAM_0001971</td>\n",
       "      <td>ISIC_0025327</td>\n",
       "      <td>nv</td>\n",
       "      <td>histo</td>\n",
       "      <td>50.0</td>\n",
       "      <td>male</td>\n",
       "      <td>back</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0025327.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.226828</td>\n",
       "      <td>1.929353</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HAM_0006476</td>\n",
       "      <td>ISIC_0031877</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>45.0</td>\n",
       "      <td>male</td>\n",
       "      <td>back</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0031877.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.957106</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HAM_0001806</td>\n",
       "      <td>ISIC_0026920</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>50.0</td>\n",
       "      <td>female</td>\n",
       "      <td>abdomen</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0026920.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HAM_0001845</td>\n",
       "      <td>ISIC_0026462</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>55.0</td>\n",
       "      <td>female</td>\n",
       "      <td>trunk</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0026462.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>5.985096</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.274430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HAM_0004963</td>\n",
       "      <td>ISIC_0033071</td>\n",
       "      <td>nv</td>\n",
       "      <td>histo</td>\n",
       "      <td>45.0</td>\n",
       "      <td>male</td>\n",
       "      <td>upper extremity</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0033071.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.370028</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 523 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     lesion_id      image_id  dx    dx_type   age     sex     localization  \\\n",
       "0  HAM_0001971  ISIC_0025327  nv      histo  50.0    male             back   \n",
       "1  HAM_0006476  ISIC_0031877  nv  follow_up  45.0    male             back   \n",
       "2  HAM_0001806  ISIC_0026920  nv  follow_up  50.0  female          abdomen   \n",
       "3  HAM_0001845  ISIC_0026462  nv  follow_up  55.0  female            trunk   \n",
       "4  HAM_0004963  ISIC_0033071  nv      histo  45.0    male  upper extremity   \n",
       "\n",
       "                                              path         cell_type  \\\n",
       "0  ./input/HAM10000_images_part_1/ISIC_0025327.jpg  Melanocytic nevi   \n",
       "1  ./input/HAM10000_images_part_1/ISIC_0031877.jpg  Melanocytic nevi   \n",
       "2  ./input/HAM10000_images_part_1/ISIC_0026920.jpg  Melanocytic nevi   \n",
       "3  ./input/HAM10000_images_part_1/ISIC_0026462.jpg  Melanocytic nevi   \n",
       "4  ./input/HAM10000_images_part_1/ISIC_0033071.jpg  Melanocytic nevi   \n",
       "\n",
       "   cell_type_idx  ...       502  503  504  505       506       507       508  \\\n",
       "0              4  ...  0.000000  0.0  0.0  0.0  0.000000  1.226828  1.929353   \n",
       "1              4  ...  0.000000  0.0  0.0  0.0  1.957106  0.000000  0.000000   \n",
       "2              4  ...  0.000000  0.0  0.0  0.0  0.000000  0.000000  0.000000   \n",
       "3              4  ...  5.985096  0.0  0.0  0.0  3.274430  0.000000  0.000000   \n",
       "4              4  ...  0.370028  0.0  0.0  0.0  0.000000  0.000000  0.000000   \n",
       "\n",
       "   509  510  511  \n",
       "0  0.0  0.0  0.0  \n",
       "1  0.0  0.0  0.0  \n",
       "2  0.0  0.0  0.0  \n",
       "3  0.0  0.0  0.0  \n",
       "4  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 523 columns]"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "complete_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "damaged-elephant",
   "metadata": {},
   "source": [
    "## Save the original and new model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "seven-corner",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_json = model_a.to_json()\n",
    "with open(\"model_a.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "requested-electricity",
   "metadata": {},
   "outputs": [],
   "source": [
    "completemodel_json = model2.to_json()\n",
    "with open(\"completemodel1.json\", \"w\") as json_file:\n",
    "    json_file.write(completemodel_json)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "authentic-pavilion",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved model to disk\n"
     ]
    }
   ],
   "source": [
    "model_a.save_weights(\"model_a_weights.h5\")\n",
    "print(\"Saved model to disk\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "sudden-cooperation",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved model to disk\n"
     ]
    }
   ],
   "source": [
    "model2.save_weights(\"completeweights.h5\")\n",
    "print(\"Saved model to disk\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "accurate-neighbor",
   "metadata": {},
   "source": [
    "## New dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "desirable-coalition",
   "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>lesion_id</th>\n",
       "      <th>image_id</th>\n",
       "      <th>dx</th>\n",
       "      <th>dx_type</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>localization</th>\n",
       "      <th>path</th>\n",
       "      <th>cell_type</th>\n",
       "      <th>cell_type_idx</th>\n",
       "      <th>...</th>\n",
       "      <th>502</th>\n",
       "      <th>503</th>\n",
       "      <th>504</th>\n",
       "      <th>505</th>\n",
       "      <th>506</th>\n",
       "      <th>507</th>\n",
       "      <th>508</th>\n",
       "      <th>509</th>\n",
       "      <th>510</th>\n",
       "      <th>511</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HAM_0001971</td>\n",
       "      <td>ISIC_0025327</td>\n",
       "      <td>nv</td>\n",
       "      <td>histo</td>\n",
       "      <td>50.0</td>\n",
       "      <td>male</td>\n",
       "      <td>back</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0025327.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.226828</td>\n",
       "      <td>1.929353</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HAM_0006476</td>\n",
       "      <td>ISIC_0031877</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>45.0</td>\n",
       "      <td>male</td>\n",
       "      <td>back</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0031877.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.957106</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HAM_0001806</td>\n",
       "      <td>ISIC_0026920</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>50.0</td>\n",
       "      <td>female</td>\n",
       "      <td>abdomen</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0026920.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HAM_0001845</td>\n",
       "      <td>ISIC_0026462</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>55.0</td>\n",
       "      <td>female</td>\n",
       "      <td>trunk</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0026462.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>5.985096</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.274430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HAM_0004963</td>\n",
       "      <td>ISIC_0033071</td>\n",
       "      <td>nv</td>\n",
       "      <td>histo</td>\n",
       "      <td>45.0</td>\n",
       "      <td>male</td>\n",
       "      <td>upper extremity</td>\n",
       "      <td>./input/HAM10000_images_part_1/ISIC_0033071.jpg</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.370028</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 523 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     lesion_id      image_id  dx    dx_type   age     sex     localization  \\\n",
       "0  HAM_0001971  ISIC_0025327  nv      histo  50.0    male             back   \n",
       "1  HAM_0006476  ISIC_0031877  nv  follow_up  45.0    male             back   \n",
       "2  HAM_0001806  ISIC_0026920  nv  follow_up  50.0  female          abdomen   \n",
       "3  HAM_0001845  ISIC_0026462  nv  follow_up  55.0  female            trunk   \n",
       "4  HAM_0004963  ISIC_0033071  nv      histo  45.0    male  upper extremity   \n",
       "\n",
       "                                              path         cell_type  \\\n",
       "0  ./input/HAM10000_images_part_1/ISIC_0025327.jpg  Melanocytic nevi   \n",
       "1  ./input/HAM10000_images_part_1/ISIC_0031877.jpg  Melanocytic nevi   \n",
       "2  ./input/HAM10000_images_part_1/ISIC_0026920.jpg  Melanocytic nevi   \n",
       "3  ./input/HAM10000_images_part_1/ISIC_0026462.jpg  Melanocytic nevi   \n",
       "4  ./input/HAM10000_images_part_1/ISIC_0033071.jpg  Melanocytic nevi   \n",
       "\n",
       "   cell_type_idx  ...       502  503  504  505       506       507       508  \\\n",
       "0              4  ...  0.000000  0.0  0.0  0.0  0.000000  1.226828  1.929353   \n",
       "1              4  ...  0.000000  0.0  0.0  0.0  1.957106  0.000000  0.000000   \n",
       "2              4  ...  0.000000  0.0  0.0  0.0  0.000000  0.000000  0.000000   \n",
       "3              4  ...  5.985096  0.0  0.0  0.0  3.274430  0.000000  0.000000   \n",
       "4              4  ...  0.370028  0.0  0.0  0.0  0.000000  0.000000  0.000000   \n",
       "\n",
       "   509  510  511  \n",
       "0  0.0  0.0  0.0  \n",
       "1  0.0  0.0  0.0  \n",
       "2  0.0  0.0  0.0  \n",
       "3  0.0  0.0  0.0  \n",
       "4  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 523 columns]"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# This dataset has original text data with the extracted features merged\n",
    "\n",
    "complete_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "younger-hungary",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "back               2192\n",
       "lower extremity    2077\n",
       "trunk              1403\n",
       "upper extremity    1116\n",
       "abdomen            1021\n",
       "face                743\n",
       "chest               405\n",
       "foot                317\n",
       "unknown             187\n",
       "neck                168\n",
       "scalp               128\n",
       "hand                 90\n",
       "ear                  56\n",
       "genital              48\n",
       "acral                 7\n",
       "Name: localization, dtype: int64"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "complete_data['localization'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "indonesian-vatican",
   "metadata": {},
   "outputs": [],
   "source": [
    "# The text cannot be processed as strings, so it will be one hot encoded"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "spare-multimedia",
   "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>confocal</th>\n",
       "      <th>consensus</th>\n",
       "      <th>follow_up</th>\n",
       "      <th>histo</th>\n",
       "      <th>lesion_id</th>\n",
       "      <th>image_id</th>\n",
       "      <th>dx</th>\n",
       "      <th>dx_type</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>...</th>\n",
       "      <th>502</th>\n",
       "      <th>503</th>\n",
       "      <th>504</th>\n",
       "      <th>505</th>\n",
       "      <th>506</th>\n",
       "      <th>507</th>\n",
       "      <th>508</th>\n",
       "      <th>509</th>\n",
       "      <th>510</th>\n",
       "      <th>511</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>HAM_0001971</td>\n",
       "      <td>ISIC_0025327</td>\n",
       "      <td>nv</td>\n",
       "      <td>histo</td>\n",
       "      <td>50.0</td>\n",
       "      <td>male</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.226828</td>\n",
       "      <td>1.929353</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>HAM_0006476</td>\n",
       "      <td>ISIC_0031877</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>45.0</td>\n",
       "      <td>male</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.957106</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>HAM_0001806</td>\n",
       "      <td>ISIC_0026920</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>50.0</td>\n",
       "      <td>female</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>HAM_0001845</td>\n",
       "      <td>ISIC_0026462</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>55.0</td>\n",
       "      <td>female</td>\n",
       "      <td>...</td>\n",
       "      <td>5.985096</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.274430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>HAM_0004963</td>\n",
       "      <td>ISIC_0033071</td>\n",
       "      <td>nv</td>\n",
       "      <td>histo</td>\n",
       "      <td>45.0</td>\n",
       "      <td>male</td>\n",
       "      <td>...</td>\n",
       "      <td>0.370028</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 527 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   confocal  consensus  follow_up  histo    lesion_id      image_id  dx  \\\n",
       "0         0          0          0      1  HAM_0001971  ISIC_0025327  nv   \n",
       "1         0          0          1      0  HAM_0006476  ISIC_0031877  nv   \n",
       "2         0          0          1      0  HAM_0001806  ISIC_0026920  nv   \n",
       "3         0          0          1      0  HAM_0001845  ISIC_0026462  nv   \n",
       "4         0          0          0      1  HAM_0004963  ISIC_0033071  nv   \n",
       "\n",
       "     dx_type   age     sex  ...       502  503  504  505       506       507  \\\n",
       "0      histo  50.0    male  ...  0.000000  0.0  0.0  0.0  0.000000  1.226828   \n",
       "1  follow_up  45.0    male  ...  0.000000  0.0  0.0  0.0  1.957106  0.000000   \n",
       "2  follow_up  50.0  female  ...  0.000000  0.0  0.0  0.0  0.000000  0.000000   \n",
       "3  follow_up  55.0  female  ...  5.985096  0.0  0.0  0.0  3.274430  0.000000   \n",
       "4      histo  45.0    male  ...  0.370028  0.0  0.0  0.0  0.000000  0.000000   \n",
       "\n",
       "        508  509  510  511  \n",
       "0  1.929353  0.0  0.0  0.0  \n",
       "1  0.000000  0.0  0.0  0.0  \n",
       "2  0.000000  0.0  0.0  0.0  \n",
       "3  0.000000  0.0  0.0  0.0  \n",
       "4  0.000000  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 527 columns]"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dxtype_df = pd.get_dummies(complete_data['dx_type'], drop_first = False)\n",
    "complete_data = pd.concat([dxtype_df, complete_data], axis = 1)\n",
    "\n",
    "# complete_data.drop(['dx_type'], axis =1 , inplace = True)\n",
    "complete_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "built-worthy",
   "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>abdomen</th>\n",
       "      <th>acral</th>\n",
       "      <th>back</th>\n",
       "      <th>chest</th>\n",
       "      <th>ear</th>\n",
       "      <th>face</th>\n",
       "      <th>foot</th>\n",
       "      <th>genital</th>\n",
       "      <th>hand</th>\n",
       "      <th>lower extremity</th>\n",
       "      <th>...</th>\n",
       "      <th>502</th>\n",
       "      <th>503</th>\n",
       "      <th>504</th>\n",
       "      <th>505</th>\n",
       "      <th>506</th>\n",
       "      <th>507</th>\n",
       "      <th>508</th>\n",
       "      <th>509</th>\n",
       "      <th>510</th>\n",
       "      <th>511</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.226828</td>\n",
       "      <td>1.929353</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.957106</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.985096</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.274430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.370028</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 542 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   abdomen  acral  back  chest  ear  face  foot  genital  hand  \\\n",
       "0        0      0     1      0    0     0     0        0     0   \n",
       "1        0      0     1      0    0     0     0        0     0   \n",
       "2        1      0     0      0    0     0     0        0     0   \n",
       "3        0      0     0      0    0     0     0        0     0   \n",
       "4        0      0     0      0    0     0     0        0     0   \n",
       "\n",
       "   lower extremity  ...       502  503  504  505       506       507  \\\n",
       "0                0  ...  0.000000  0.0  0.0  0.0  0.000000  1.226828   \n",
       "1                0  ...  0.000000  0.0  0.0  0.0  1.957106  0.000000   \n",
       "2                0  ...  0.000000  0.0  0.0  0.0  0.000000  0.000000   \n",
       "3                0  ...  5.985096  0.0  0.0  0.0  3.274430  0.000000   \n",
       "4                0  ...  0.370028  0.0  0.0  0.0  0.000000  0.000000   \n",
       "\n",
       "        508  509  510  511  \n",
       "0  1.929353  0.0  0.0  0.0  \n",
       "1  0.000000  0.0  0.0  0.0  \n",
       "2  0.000000  0.0  0.0  0.0  \n",
       "3  0.000000  0.0  0.0  0.0  \n",
       "4  0.000000  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 542 columns]"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "localization_df = pd.get_dummies(complete_data['localization'], drop_first = False)\n",
    "complete_data = pd.concat([localization_df, complete_data], axis = 1)\n",
    "\n",
    "# complete_data.drop(['dx_type'], axis = 1, inplace = True)\n",
    "complete_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "measured-consciousness",
   "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>female</th>\n",
       "      <th>male</th>\n",
       "      <th>abdomen</th>\n",
       "      <th>acral</th>\n",
       "      <th>back</th>\n",
       "      <th>chest</th>\n",
       "      <th>ear</th>\n",
       "      <th>face</th>\n",
       "      <th>foot</th>\n",
       "      <th>genital</th>\n",
       "      <th>...</th>\n",
       "      <th>502</th>\n",
       "      <th>503</th>\n",
       "      <th>504</th>\n",
       "      <th>505</th>\n",
       "      <th>506</th>\n",
       "      <th>507</th>\n",
       "      <th>508</th>\n",
       "      <th>509</th>\n",
       "      <th>510</th>\n",
       "      <th>511</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.226828</td>\n",
       "      <td>1.929353</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.957106</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.985096</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.274430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.370028</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 544 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   female  male  abdomen  acral  back  chest  ear  face  foot  genital  ...  \\\n",
       "0       0     1        0      0     1      0    0     0     0        0  ...   \n",
       "1       0     1        0      0     1      0    0     0     0        0  ...   \n",
       "2       1     0        1      0     0      0    0     0     0        0  ...   \n",
       "3       1     0        0      0     0      0    0     0     0        0  ...   \n",
       "4       0     1        0      0     0      0    0     0     0        0  ...   \n",
       "\n",
       "        502  503  504  505       506       507       508  509  510  511  \n",
       "0  0.000000  0.0  0.0  0.0  0.000000  1.226828  1.929353  0.0  0.0  0.0  \n",
       "1  0.000000  0.0  0.0  0.0  1.957106  0.000000  0.000000  0.0  0.0  0.0  \n",
       "2  0.000000  0.0  0.0  0.0  0.000000  0.000000  0.000000  0.0  0.0  0.0  \n",
       "3  5.985096  0.0  0.0  0.0  3.274430  0.000000  0.000000  0.0  0.0  0.0  \n",
       "4  0.370028  0.0  0.0  0.0  0.000000  0.000000  0.000000  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 544 columns]"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sex_df = pd.get_dummies(complete_data['sex'], drop_first = False)\n",
    "sex_df.drop(['unknown'], axis = 1, inplace = True)\n",
    "complete_data = pd.concat([sex_df, complete_data], axis = 1)\n",
    "\n",
    "# complete_data.drop(['dx_type'], axis = 1, inplace = True)\n",
    "complete_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "deluxe-plymouth",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9958, 544)"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "complete_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "immediate-cemetery",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_labels = complete_data.drop(['lesion_id','image_id','dx_type','dx','path','cell_type','cell_type_idx',\n",
    "                               'sex','path','localization','image'], axis = 1, inplace = False)\n",
    "y_label = complete_data['cell_type_idx']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "illegal-motorcycle",
   "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>female</th>\n",
       "      <th>male</th>\n",
       "      <th>abdomen</th>\n",
       "      <th>acral</th>\n",
       "      <th>back</th>\n",
       "      <th>chest</th>\n",
       "      <th>ear</th>\n",
       "      <th>face</th>\n",
       "      <th>foot</th>\n",
       "      <th>genital</th>\n",
       "      <th>...</th>\n",
       "      <th>502</th>\n",
       "      <th>503</th>\n",
       "      <th>504</th>\n",
       "      <th>505</th>\n",
       "      <th>506</th>\n",
       "      <th>507</th>\n",
       "      <th>508</th>\n",
       "      <th>509</th>\n",
       "      <th>510</th>\n",
       "      <th>511</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.226828</td>\n",
       "      <td>1.929353</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.957106</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.985096</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.274430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.370028</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 534 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   female  male  abdomen  acral  back  chest  ear  face  foot  genital  ...  \\\n",
       "0       0     1        0      0     1      0    0     0     0        0  ...   \n",
       "1       0     1        0      0     1      0    0     0     0        0  ...   \n",
       "2       1     0        1      0     0      0    0     0     0        0  ...   \n",
       "3       1     0        0      0     0      0    0     0     0        0  ...   \n",
       "4       0     1        0      0     0      0    0     0     0        0  ...   \n",
       "\n",
       "        502  503  504  505       506       507       508  509  510  511  \n",
       "0  0.000000  0.0  0.0  0.0  0.000000  1.226828  1.929353  0.0  0.0  0.0  \n",
       "1  0.000000  0.0  0.0  0.0  1.957106  0.000000  0.000000  0.0  0.0  0.0  \n",
       "2  0.000000  0.0  0.0  0.0  0.000000  0.000000  0.000000  0.0  0.0  0.0  \n",
       "3  5.985096  0.0  0.0  0.0  3.274430  0.000000  0.000000  0.0  0.0  0.0  \n",
       "4  0.370028  0.0  0.0  0.0  0.000000  0.000000  0.000000  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 534 columns]"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_labels.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "automotive-while",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9958, 534)"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_labels.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "asian-sperm",
   "metadata": {},
   "source": [
    "## Saving the newly created dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "instructional-circumstances",
   "metadata": {},
   "outputs": [],
   "source": [
    "complete_data.to_csv('skin_data_v2.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "superb-tooth",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1992, 75, 75, 3)\n",
      "(1992, 7)\n"
     ]
    }
   ],
   "source": [
    "print(x_test.shape)\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "norwegian-skiing",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "[0. 0. 0. 0. 1. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "# Must convert back to label encoding\n",
    "\n",
    "print(type(y_test))\n",
    "print(y_test[10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "blocked-cancellation",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test2 = np.array(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "painted-decade",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0., ..., 1., 0., 0.],\n",
       "       [0., 0., 0., ..., 1., 0., 0.],\n",
       "       [0., 0., 1., ..., 0., 0., 0.],\n",
       "       ...,\n",
       "       [0., 0., 1., ..., 0., 0., 0.],\n",
       "       [0., 0., 0., ..., 1., 0., 0.],\n",
       "       [0., 0., 0., ..., 1., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "animated-penetration",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test2 = np.where(y_test2)[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "suspended-murray",
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model2.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "serious-office",
   "metadata": {},
   "outputs": [],
   "source": [
    "lst = []\n",
    "for a in preds:\n",
    "    lst.append(np.argmax(a))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "stupid-moore",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import classification_report"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "unlimited-access",
   "metadata": {},
   "source": [
    "## CNN Model evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "quarterly-doubt",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  17,    5,    4,    0,    2,    5,    0],\n",
       "       [   8,   32,   14,    0,    6,   12,    4],\n",
       "       [  35,   29,  110,    3,   64,   64,    0],\n",
       "       [   0,    3,    0,    3,    3,    0,    0],\n",
       "       [   0,    2,    4,    0,   15,   38,    1],\n",
       "       [  18,   22,   82,   11, 1207,  122,   19],\n",
       "       [   0,    0,    1,    0,   12,    1,   14]])"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(lst, y_test2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "wound-following",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.1571285140562249"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(lst, y_test2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "final-killing",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.52      0.22      0.31        78\n",
      "           1       0.42      0.34      0.38        93\n",
      "           2       0.36      0.51      0.42       215\n",
      "           3       0.33      0.18      0.23        17\n",
      "           4       0.25      0.01      0.02      1309\n",
      "           5       0.08      0.50      0.14       242\n",
      "           6       0.50      0.37      0.42        38\n",
      "\n",
      "    accuracy                           0.16      1992\n",
      "   macro avg       0.35      0.30      0.28      1992\n",
      "weighted avg       0.27      0.16      0.12      1992\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test2, lst))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "conceptual-botswana",
   "metadata": {},
   "source": [
    "## Importing the new dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "structural-congo",
   "metadata": {},
   "outputs": [],
   "source": [
    "skin_df = pd.read_csv('skin_data_v2.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "worth-smell",
   "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>female</th>\n",
       "      <th>male</th>\n",
       "      <th>abdomen</th>\n",
       "      <th>acral</th>\n",
       "      <th>back</th>\n",
       "      <th>chest</th>\n",
       "      <th>ear</th>\n",
       "      <th>face</th>\n",
       "      <th>foot</th>\n",
       "      <th>genital</th>\n",
       "      <th>...</th>\n",
       "      <th>502</th>\n",
       "      <th>503</th>\n",
       "      <th>504</th>\n",
       "      <th>505</th>\n",
       "      <th>506</th>\n",
       "      <th>507</th>\n",
       "      <th>508</th>\n",
       "      <th>509</th>\n",
       "      <th>510</th>\n",
       "      <th>511</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.226827</td>\n",
       "      <td>1.929353</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.957106</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.985095</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.274430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.370028</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 544 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   female  male  abdomen  acral  back  chest  ear  face  foot  genital  ...  \\\n",
       "0       0     1        0      0     1      0    0     0     0        0  ...   \n",
       "1       0     1        0      0     1      0    0     0     0        0  ...   \n",
       "2       1     0        1      0     0      0    0     0     0        0  ...   \n",
       "3       1     0        0      0     0      0    0     0     0        0  ...   \n",
       "4       0     1        0      0     0      0    0     0     0        0  ...   \n",
       "\n",
       "        502  503  504  505       506       507       508  509  510  511  \n",
       "0  0.000000  0.0  0.0  0.0  0.000000  1.226827  1.929353  0.0  0.0  0.0  \n",
       "1  0.000000  0.0  0.0  0.0  1.957106  0.000000  0.000000  0.0  0.0  0.0  \n",
       "2  0.000000  0.0  0.0  0.0  0.000000  0.000000  0.000000  0.0  0.0  0.0  \n",
       "3  5.985095  0.0  0.0  0.0  3.274430  0.000000  0.000000  0.0  0.0  0.0  \n",
       "4  0.370028  0.0  0.0  0.0  0.000000  0.000000  0.000000  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 544 columns]"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "skin_df.drop(['Unnamed: 0'],axis=1,inplace=True)\n",
    "skin_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "ideal-combination",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pycaret in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (2.3.0)\n",
      "Requirement already satisfied: scikit-plot in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.3.7)\n",
      "Requirement already satisfied: gensim in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (4.0.0)\n",
      "Requirement already satisfied: ipywidgets in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (7.6.3)\n",
      "Requirement already satisfied: imbalanced-learn>=0.7.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.7.0)\n",
      "Requirement already satisfied: pandas-profiling>=2.8.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (2.11.0)\n",
      "Requirement already satisfied: IPython in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (7.19.0)\n",
      "Requirement already satisfied: kmodes>=0.10.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.11.0)\n",
      "Requirement already satisfied: matplotlib in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (3.3.2)\n",
      "Requirement already satisfied: yellowbrick>=1.0.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (1.3.post1)\n",
      "Requirement already satisfied: scipy<=1.5.4 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (1.5.2)\n",
      "Requirement already satisfied: plotly>=4.4.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (4.14.3)\n",
      "Requirement already satisfied: cufflinks>=0.17.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.17.3)\n",
      "Requirement already satisfied: mlxtend in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.18.0)\n",
      "Requirement already satisfied: textblob in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.15.3)\n",
      "Requirement already satisfied: pandas in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (1.2.0)\n",
      "Requirement already satisfied: pyLDAvis in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (3.2.2)\n",
      "Requirement already satisfied: joblib in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (1.0.0)\n",
      "Requirement already satisfied: numpy==1.19.5 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (1.19.5)\n",
      "Requirement already satisfied: mlflow in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (1.14.1)\n",
      "Requirement already satisfied: umap-learn in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.5.1)\n",
      "Requirement already satisfied: seaborn in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.11.1)\n",
      "Requirement already satisfied: wordcloud in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (1.8.1)\n",
      "Requirement already satisfied: spacy<2.4.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (2.3.5)\n",
      "Requirement already satisfied: Boruta in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.3)\n",
      "Requirement already satisfied: pyod in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.8.7)\n",
      "Requirement already satisfied: nltk in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (3.5)\n",
      "Requirement already satisfied: scikit-learn==0.23.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (0.23.2)\n",
      "Requirement already satisfied: lightgbm>=2.3.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pycaret) (3.2.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from scikit-learn==0.23.2->pycaret) (2.1.0)\n",
      "Requirement already satisfied: colorlover>=0.2.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from cufflinks>=0.17.0->pycaret) (0.3.0)\n",
      "Requirement already satisfied: setuptools>=34.4.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from cufflinks>=0.17.0->pycaret) (51.1.2.post20210112)\n",
      "Requirement already satisfied: six>=1.9.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from cufflinks>=0.17.0->pycaret) (1.15.0)\n",
      "Requirement already satisfied: prompt-toolkit!=3.0.0,!=3.0.1,<3.1.0,>=2.0.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (3.0.8)\n",
      "Requirement already satisfied: backcall in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (0.2.0)\n",
      "Requirement already satisfied: appnope in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (0.1.2)\n",
      "Requirement already satisfied: jedi>=0.10 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (0.14.1)\n",
      "Requirement already satisfied: pygments in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (2.7.4)\n",
      "Requirement already satisfied: pickleshare in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (0.7.5)\n",
      "Requirement already satisfied: decorator in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (4.4.2)\n",
      "Requirement already satisfied: traitlets>=4.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (5.0.5)\n",
      "Requirement already satisfied: pexpect>4.3 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from IPython->pycaret) (4.8.0)\n",
      "Requirement already satisfied: widgetsnbextension~=3.5.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from ipywidgets->pycaret) (3.5.1)\n",
      "Requirement already satisfied: jupyterlab-widgets>=1.0.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from ipywidgets->pycaret) (1.0.0)\n",
      "Requirement already satisfied: ipykernel>=4.5.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from ipywidgets->pycaret) (5.3.4)\n",
      "Requirement already satisfied: nbformat>=4.2.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from ipywidgets->pycaret) (5.1.2)\n",
      "Requirement already satisfied: tornado>=4.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from ipykernel>=4.5.1->ipywidgets->pycaret) (6.1)\n",
      "Requirement already satisfied: jupyter-client in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from ipykernel>=4.5.1->ipywidgets->pycaret) (6.1.7)\n",
      "Requirement already satisfied: parso>=0.5.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from jedi>=0.10->IPython->pycaret) (0.5.2)\n",
      "Requirement already satisfied: wheel in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from lightgbm>=2.3.1->pycaret) (0.36.2)\n",
      "Requirement already satisfied: jupyter-core in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbformat>=4.2.0->ipywidgets->pycaret) (4.7.0)\n",
      "Requirement already satisfied: ipython-genutils in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbformat>=4.2.0->ipywidgets->pycaret) (0.2.0)\n",
      "Requirement already satisfied: jsonschema!=2.5.0,>=2.4 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbformat>=4.2.0->ipywidgets->pycaret) (3.2.0)\n",
      "Requirement already satisfied: importlib-metadata in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->pycaret) (2.0.0)\n",
      "Requirement already satisfied: pyrsistent>=0.14.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->pycaret) (0.17.3)\n",
      "Requirement already satisfied: attrs>=17.4.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->pycaret) (20.3.0)\n",
      "Requirement already satisfied: python-dateutil>=2.7.3 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas->pycaret) (2.8.1)\n",
      "Requirement already satisfied: pytz>=2017.3 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas->pycaret) (2020.5)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: confuse>=1.0.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (1.4.0)\n",
      "Requirement already satisfied: requests>=2.24.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (2.25.1)\n",
      "Requirement already satisfied: htmlmin>=0.1.12 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (0.1.12)\n",
      "Requirement already satisfied: visions[type_image_path]==0.6.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (0.6.0)\n",
      "Requirement already satisfied: phik>=0.10.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (0.11.2)\n",
      "Requirement already satisfied: missingno>=0.4.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (0.4.2)\n",
      "Requirement already satisfied: tangled-up-in-unicode>=0.0.6 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (0.0.7)\n",
      "Requirement already satisfied: tqdm>=4.48.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (4.55.1)\n",
      "Requirement already satisfied: jinja2>=2.11.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pandas-profiling>=2.8.0->pycaret) (2.11.2)\n",
      "Requirement already satisfied: networkx>=2.4 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from visions[type_image_path]==0.6.0->pandas-profiling>=2.8.0->pycaret) (2.5)\n",
      "Requirement already satisfied: Pillow in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from visions[type_image_path]==0.6.0->pandas-profiling>=2.8.0->pycaret) (7.0.0)\n",
      "Requirement already satisfied: imagehash in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from visions[type_image_path]==0.6.0->pandas-profiling>=2.8.0->pycaret) (4.2.0)\n",
      "Requirement already satisfied: pyyaml in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from confuse>=1.0.0->pandas-profiling>=2.8.0->pycaret) (5.3.1)\n",
      "Requirement already satisfied: MarkupSafe>=0.23 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from jinja2>=2.11.1->pandas-profiling>=2.8.0->pycaret) (1.1.1)\n",
      "Requirement already satisfied: certifi>=2020.06.20 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from matplotlib->pycaret) (2020.12.5)\n",
      "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.3 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from matplotlib->pycaret) (2.4.7)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from matplotlib->pycaret) (1.3.0)\n",
      "Requirement already satisfied: cycler>=0.10 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from matplotlib->pycaret) (0.10.0)\n",
      "Requirement already satisfied: ptyprocess>=0.5 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pexpect>4.3->IPython->pycaret) (0.7.0)\n",
      "Requirement already satisfied: retrying>=1.3.3 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from plotly>=4.4.1->pycaret) (1.3.3)\n",
      "Requirement already satisfied: wcwidth in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from prompt-toolkit!=3.0.0,!=3.0.1,<3.1.0,>=2.0.0->IPython->pycaret) (0.2.5)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from requests>=2.24.0->pandas-profiling>=2.8.0->pycaret) (2.10)\n",
      "Requirement already satisfied: chardet<5,>=3.0.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from requests>=2.24.0->pandas-profiling>=2.8.0->pycaret) (4.0.0)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from requests>=2.24.0->pandas-profiling>=2.8.0->pycaret) (1.26.2)\n",
      "Requirement already satisfied: wasabi<1.1.0,>=0.4.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (0.8.2)\n",
      "Requirement already satisfied: plac<1.2.0,>=0.9.6 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (1.1.3)\n",
      "Requirement already satisfied: catalogue<1.1.0,>=0.0.7 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (1.0.0)\n",
      "Requirement already satisfied: thinc<7.5.0,>=7.4.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (7.4.5)\n",
      "Requirement already satisfied: cymem<2.1.0,>=2.0.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (2.0.5)\n",
      "Requirement already satisfied: blis<0.8.0,>=0.4.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (0.7.4)\n",
      "Requirement already satisfied: murmurhash<1.1.0,>=0.28.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (1.0.5)\n",
      "Requirement already satisfied: preshed<3.1.0,>=3.0.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (3.0.5)\n",
      "Requirement already satisfied: srsly<1.1.0,>=1.0.2 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from spacy<2.4.0->pycaret) (1.0.5)\n",
      "Requirement already satisfied: zipp>=0.5 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from importlib-metadata->jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->pycaret) (3.4.0)\n",
      "Requirement already satisfied: notebook>=4.4.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from widgetsnbextension~=3.5.0->ipywidgets->pycaret) (6.1.6)\n",
      "Requirement already satisfied: Send2Trash in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (1.5.0)\n",
      "Requirement already satisfied: argon2-cffi in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (20.1.0)\n",
      "Requirement already satisfied: prometheus-client in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.9.0)\n",
      "Requirement already satisfied: nbconvert in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (6.0.7)\n",
      "Requirement already satisfied: terminado>=0.8.3 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.9.2)\n",
      "Requirement already satisfied: pyzmq>=17 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (20.0.0)\n",
      "Requirement already satisfied: cffi>=1.0.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from argon2-cffi->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (1.14.4)\n",
      "Requirement already satisfied: pycparser in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from cffi>=1.0.0->argon2-cffi->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (2.20)\n",
      "Requirement already satisfied: smart-open>=1.8.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from gensim->pycaret) (4.2.0)\n",
      "Requirement already satisfied: PyWavelets in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from imagehash->visions[type_image_path]==0.6.0->pandas-profiling>=2.8.0->pycaret) (1.1.1)\n",
      "Requirement already satisfied: gitpython>=2.1.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (3.1.14)\n",
      "Requirement already satisfied: entrypoints in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (0.3)\n",
      "Requirement already satisfied: prometheus-flask-exporter in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (0.18.1)\n",
      "Requirement already satisfied: databricks-cli>=0.8.7 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (0.14.3)\n",
      "Requirement already satisfied: sqlalchemy in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (1.3.21)\n",
      "Requirement already satisfied: click>=7.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (7.1.2)\n",
      "Requirement already satisfied: protobuf>=3.6.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (3.13.0)\n",
      "Requirement already satisfied: Flask in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (1.1.2)\n",
      "Requirement already satisfied: cloudpickle in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (1.6.0)\n",
      "Requirement already satisfied: alembic<=1.4.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (1.4.1)\n",
      "Requirement already satisfied: docker>=4.0.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (4.4.4)\n",
      "Requirement already satisfied: querystring-parser in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (1.2.4)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: gunicorn in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (20.0.4)\n",
      "Requirement already satisfied: sqlparse>=0.3.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from mlflow->pycaret) (0.4.1)\n",
      "Requirement already satisfied: Mako in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from alembic<=1.4.1->mlflow->pycaret) (1.1.4)\n",
      "Requirement already satisfied: python-editor>=0.3 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from alembic<=1.4.1->mlflow->pycaret) (1.0.4)\n",
      "Requirement already satisfied: tabulate>=0.7.7 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from databricks-cli>=0.8.7->mlflow->pycaret) (0.8.9)\n",
      "Requirement already satisfied: websocket-client>=0.32.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from docker>=4.0.0->mlflow->pycaret) (0.58.0)\n",
      "Requirement already satisfied: gitdb<5,>=4.0.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from gitpython>=2.1.0->mlflow->pycaret) (4.0.7)\n",
      "Requirement already satisfied: smmap<5,>=3.0.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from gitdb<5,>=4.0.1->gitpython>=2.1.0->mlflow->pycaret) (4.0.0)\n",
      "Requirement already satisfied: itsdangerous>=0.24 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from Flask->mlflow->pycaret) (1.1.0)\n",
      "Requirement already satisfied: Werkzeug>=0.15 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from Flask->mlflow->pycaret) (1.0.1)\n",
      "Requirement already satisfied: mistune<2,>=0.8.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.8.4)\n",
      "Requirement already satisfied: bleach in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (3.2.1)\n",
      "Requirement already satisfied: defusedxml in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.6.0)\n",
      "Requirement already satisfied: nbclient<0.6.0,>=0.5.0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.5.1)\n",
      "Requirement already satisfied: jupyterlab-pygments in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.1.2)\n",
      "Requirement already satisfied: pandocfilters>=1.4.1 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (1.4.3)\n",
      "Requirement already satisfied: testpath in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.4.4)\n",
      "Requirement already satisfied: async-generator in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbclient<0.6.0,>=0.5.0->nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (1.10)\n",
      "Requirement already satisfied: nest-asyncio in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nbclient<0.6.0,>=0.5.0->nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (1.4.3)\n",
      "Requirement already satisfied: webencodings in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from bleach->nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (0.5.1)\n",
      "Requirement already satisfied: packaging in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from bleach->nbconvert->notebook>=4.4.1->widgetsnbextension~=3.5.0->ipywidgets->pycaret) (20.8)\n",
      "Requirement already satisfied: regex in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from nltk->pycaret) (2020.11.13)\n",
      "Requirement already satisfied: funcy in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pyLDAvis->pycaret) (1.15)\n",
      "Requirement already satisfied: numexpr in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pyLDAvis->pycaret) (2.7.2)\n",
      "Requirement already satisfied: future in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pyLDAvis->pycaret) (0.18.2)\n",
      "Requirement already satisfied: statsmodels in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pyod->pycaret) (0.12.1)\n",
      "Requirement already satisfied: numba>=0.35 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from pyod->pycaret) (0.51.2)\n",
      "Requirement already satisfied: llvmlite<0.35,>=0.34.0.dev0 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from numba>=0.35->pyod->pycaret) (0.34.0)\n",
      "Requirement already satisfied: patsy>=0.5 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from statsmodels->pyod->pycaret) (0.5.1)\n",
      "Requirement already satisfied: pynndescent>=0.5 in /Users/sparshbohra/opt/anaconda3/lib/python3.7/site-packages (from umap-learn->pycaret) (0.5.2)\n"
     ]
    }
   ],
   "source": [
    "!pip install pycaret"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "structural-celebrity",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_labels = skin_df.drop(['lesion_id','image_id','dx','dx_type','sex','localization',\n",
    "                         'path','cell_type','cell_type_idx','image'], axis=1,inplace=False)\n",
    "y_label = skin_df['cell_type_idx']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "bulgarian-latter",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_classification = pd.concat([X_labels, y_label], axis = 1)\n",
    "data_classification.fillna(data_classification['age'].mean(), inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "offensive-pride",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "</style><table id=\"T_50c18_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Description</th>        <th class=\"col_heading level0 col1\" >Value</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_50c18_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_50c18_row0_col0\" class=\"data row0 col0\" >session_id</td>\n",
       "                        <td id=\"T_50c18_row0_col1\" class=\"data row0 col1\" >5653</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_50c18_row1_col0\" class=\"data row1 col0\" >Target</td>\n",
       "                        <td id=\"T_50c18_row1_col1\" class=\"data row1 col1\" >cell_type_idx</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_50c18_row2_col0\" class=\"data row2 col0\" >Target Type</td>\n",
       "                        <td id=\"T_50c18_row2_col1\" class=\"data row2 col1\" >Multiclass</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_50c18_row3_col0\" class=\"data row3 col0\" >Label Encoded</td>\n",
       "                        <td id=\"T_50c18_row3_col1\" class=\"data row3 col1\" >0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5, 6: 6</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_50c18_row4_col0\" class=\"data row4 col0\" >Original Data</td>\n",
       "                        <td id=\"T_50c18_row4_col1\" class=\"data row4 col1\" >(9958, 535)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row5\" class=\"row_heading level0 row5\" >5</th>\n",
       "                        <td id=\"T_50c18_row5_col0\" class=\"data row5 col0\" >Missing Values</td>\n",
       "                        <td id=\"T_50c18_row5_col1\" class=\"data row5 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row6\" class=\"row_heading level0 row6\" >6</th>\n",
       "                        <td id=\"T_50c18_row6_col0\" class=\"data row6 col0\" >Numeric Features</td>\n",
       "                        <td id=\"T_50c18_row6_col1\" class=\"data row6 col1\" >473</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row7\" class=\"row_heading level0 row7\" >7</th>\n",
       "                        <td id=\"T_50c18_row7_col0\" class=\"data row7 col0\" >Categorical Features</td>\n",
       "                        <td id=\"T_50c18_row7_col1\" class=\"data row7 col1\" >61</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row8\" class=\"row_heading level0 row8\" >8</th>\n",
       "                        <td id=\"T_50c18_row8_col0\" class=\"data row8 col0\" >Ordinal Features</td>\n",
       "                        <td id=\"T_50c18_row8_col1\" class=\"data row8 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row9\" class=\"row_heading level0 row9\" >9</th>\n",
       "                        <td id=\"T_50c18_row9_col0\" class=\"data row9 col0\" >High Cardinality Features</td>\n",
       "                        <td id=\"T_50c18_row9_col1\" class=\"data row9 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row10\" class=\"row_heading level0 row10\" >10</th>\n",
       "                        <td id=\"T_50c18_row10_col0\" class=\"data row10 col0\" >High Cardinality Method</td>\n",
       "                        <td id=\"T_50c18_row10_col1\" class=\"data row10 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row11\" class=\"row_heading level0 row11\" >11</th>\n",
       "                        <td id=\"T_50c18_row11_col0\" class=\"data row11 col0\" >Transformed Train Set</td>\n",
       "                        <td id=\"T_50c18_row11_col1\" class=\"data row11 col1\" >(6970, 530)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row12\" class=\"row_heading level0 row12\" >12</th>\n",
       "                        <td id=\"T_50c18_row12_col0\" class=\"data row12 col0\" >Transformed Test Set</td>\n",
       "                        <td id=\"T_50c18_row12_col1\" class=\"data row12 col1\" >(2988, 530)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row13\" class=\"row_heading level0 row13\" >13</th>\n",
       "                        <td id=\"T_50c18_row13_col0\" class=\"data row13 col0\" >Shuffle Train-Test</td>\n",
       "                        <td id=\"T_50c18_row13_col1\" class=\"data row13 col1\" >True</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row14\" class=\"row_heading level0 row14\" >14</th>\n",
       "                        <td id=\"T_50c18_row14_col0\" class=\"data row14 col0\" >Stratify Train-Test</td>\n",
       "                        <td id=\"T_50c18_row14_col1\" class=\"data row14 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row15\" class=\"row_heading level0 row15\" >15</th>\n",
       "                        <td id=\"T_50c18_row15_col0\" class=\"data row15 col0\" >Fold Generator</td>\n",
       "                        <td id=\"T_50c18_row15_col1\" class=\"data row15 col1\" >StratifiedKFold</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row16\" class=\"row_heading level0 row16\" >16</th>\n",
       "                        <td id=\"T_50c18_row16_col0\" class=\"data row16 col0\" >Fold Number</td>\n",
       "                        <td id=\"T_50c18_row16_col1\" class=\"data row16 col1\" >10</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row17\" class=\"row_heading level0 row17\" >17</th>\n",
       "                        <td id=\"T_50c18_row17_col0\" class=\"data row17 col0\" >CPU Jobs</td>\n",
       "                        <td id=\"T_50c18_row17_col1\" class=\"data row17 col1\" >-1</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row18\" class=\"row_heading level0 row18\" >18</th>\n",
       "                        <td id=\"T_50c18_row18_col0\" class=\"data row18 col0\" >Use GPU</td>\n",
       "                        <td id=\"T_50c18_row18_col1\" class=\"data row18 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row19\" class=\"row_heading level0 row19\" >19</th>\n",
       "                        <td id=\"T_50c18_row19_col0\" class=\"data row19 col0\" >Log Experiment</td>\n",
       "                        <td id=\"T_50c18_row19_col1\" class=\"data row19 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row20\" class=\"row_heading level0 row20\" >20</th>\n",
       "                        <td id=\"T_50c18_row20_col0\" class=\"data row20 col0\" >Experiment Name</td>\n",
       "                        <td id=\"T_50c18_row20_col1\" class=\"data row20 col1\" >clf-default-name</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row21\" class=\"row_heading level0 row21\" >21</th>\n",
       "                        <td id=\"T_50c18_row21_col0\" class=\"data row21 col0\" >USI</td>\n",
       "                        <td id=\"T_50c18_row21_col1\" class=\"data row21 col1\" >b5c7</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row22\" class=\"row_heading level0 row22\" >22</th>\n",
       "                        <td id=\"T_50c18_row22_col0\" class=\"data row22 col0\" >Imputation Type</td>\n",
       "                        <td id=\"T_50c18_row22_col1\" class=\"data row22 col1\" >simple</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row23\" class=\"row_heading level0 row23\" >23</th>\n",
       "                        <td id=\"T_50c18_row23_col0\" class=\"data row23 col0\" >Iterative Imputation Iteration</td>\n",
       "                        <td id=\"T_50c18_row23_col1\" class=\"data row23 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row24\" class=\"row_heading level0 row24\" >24</th>\n",
       "                        <td id=\"T_50c18_row24_col0\" class=\"data row24 col0\" >Numeric Imputer</td>\n",
       "                        <td id=\"T_50c18_row24_col1\" class=\"data row24 col1\" >mean</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row25\" class=\"row_heading level0 row25\" >25</th>\n",
       "                        <td id=\"T_50c18_row25_col0\" class=\"data row25 col0\" >Iterative Imputation Numeric Model</td>\n",
       "                        <td id=\"T_50c18_row25_col1\" class=\"data row25 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row26\" class=\"row_heading level0 row26\" >26</th>\n",
       "                        <td id=\"T_50c18_row26_col0\" class=\"data row26 col0\" >Categorical Imputer</td>\n",
       "                        <td id=\"T_50c18_row26_col1\" class=\"data row26 col1\" >constant</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row27\" class=\"row_heading level0 row27\" >27</th>\n",
       "                        <td id=\"T_50c18_row27_col0\" class=\"data row27 col0\" >Iterative Imputation Categorical Model</td>\n",
       "                        <td id=\"T_50c18_row27_col1\" class=\"data row27 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row28\" class=\"row_heading level0 row28\" >28</th>\n",
       "                        <td id=\"T_50c18_row28_col0\" class=\"data row28 col0\" >Unknown Categoricals Handling</td>\n",
       "                        <td id=\"T_50c18_row28_col1\" class=\"data row28 col1\" >least_frequent</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row29\" class=\"row_heading level0 row29\" >29</th>\n",
       "                        <td id=\"T_50c18_row29_col0\" class=\"data row29 col0\" >Normalize</td>\n",
       "                        <td id=\"T_50c18_row29_col1\" class=\"data row29 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row30\" class=\"row_heading level0 row30\" >30</th>\n",
       "                        <td id=\"T_50c18_row30_col0\" class=\"data row30 col0\" >Normalize Method</td>\n",
       "                        <td id=\"T_50c18_row30_col1\" class=\"data row30 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row31\" class=\"row_heading level0 row31\" >31</th>\n",
       "                        <td id=\"T_50c18_row31_col0\" class=\"data row31 col0\" >Transformation</td>\n",
       "                        <td id=\"T_50c18_row31_col1\" class=\"data row31 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row32\" class=\"row_heading level0 row32\" >32</th>\n",
       "                        <td id=\"T_50c18_row32_col0\" class=\"data row32 col0\" >Transformation Method</td>\n",
       "                        <td id=\"T_50c18_row32_col1\" class=\"data row32 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row33\" class=\"row_heading level0 row33\" >33</th>\n",
       "                        <td id=\"T_50c18_row33_col0\" class=\"data row33 col0\" >PCA</td>\n",
       "                        <td id=\"T_50c18_row33_col1\" class=\"data row33 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row34\" class=\"row_heading level0 row34\" >34</th>\n",
       "                        <td id=\"T_50c18_row34_col0\" class=\"data row34 col0\" >PCA Method</td>\n",
       "                        <td id=\"T_50c18_row34_col1\" class=\"data row34 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row35\" class=\"row_heading level0 row35\" >35</th>\n",
       "                        <td id=\"T_50c18_row35_col0\" class=\"data row35 col0\" >PCA Components</td>\n",
       "                        <td id=\"T_50c18_row35_col1\" class=\"data row35 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row36\" class=\"row_heading level0 row36\" >36</th>\n",
       "                        <td id=\"T_50c18_row36_col0\" class=\"data row36 col0\" >Ignore Low Variance</td>\n",
       "                        <td id=\"T_50c18_row36_col1\" class=\"data row36 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row37\" class=\"row_heading level0 row37\" >37</th>\n",
       "                        <td id=\"T_50c18_row37_col0\" class=\"data row37 col0\" >Combine Rare Levels</td>\n",
       "                        <td id=\"T_50c18_row37_col1\" class=\"data row37 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row38\" class=\"row_heading level0 row38\" >38</th>\n",
       "                        <td id=\"T_50c18_row38_col0\" class=\"data row38 col0\" >Rare Level Threshold</td>\n",
       "                        <td id=\"T_50c18_row38_col1\" class=\"data row38 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row39\" class=\"row_heading level0 row39\" >39</th>\n",
       "                        <td id=\"T_50c18_row39_col0\" class=\"data row39 col0\" >Numeric Binning</td>\n",
       "                        <td id=\"T_50c18_row39_col1\" class=\"data row39 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row40\" class=\"row_heading level0 row40\" >40</th>\n",
       "                        <td id=\"T_50c18_row40_col0\" class=\"data row40 col0\" >Remove Outliers</td>\n",
       "                        <td id=\"T_50c18_row40_col1\" class=\"data row40 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row41\" class=\"row_heading level0 row41\" >41</th>\n",
       "                        <td id=\"T_50c18_row41_col0\" class=\"data row41 col0\" >Outliers Threshold</td>\n",
       "                        <td id=\"T_50c18_row41_col1\" class=\"data row41 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row42\" class=\"row_heading level0 row42\" >42</th>\n",
       "                        <td id=\"T_50c18_row42_col0\" class=\"data row42 col0\" >Remove Multicollinearity</td>\n",
       "                        <td id=\"T_50c18_row42_col1\" class=\"data row42 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row43\" class=\"row_heading level0 row43\" >43</th>\n",
       "                        <td id=\"T_50c18_row43_col0\" class=\"data row43 col0\" >Multicollinearity Threshold</td>\n",
       "                        <td id=\"T_50c18_row43_col1\" class=\"data row43 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row44\" class=\"row_heading level0 row44\" >44</th>\n",
       "                        <td id=\"T_50c18_row44_col0\" class=\"data row44 col0\" >Clustering</td>\n",
       "                        <td id=\"T_50c18_row44_col1\" class=\"data row44 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row45\" class=\"row_heading level0 row45\" >45</th>\n",
       "                        <td id=\"T_50c18_row45_col0\" class=\"data row45 col0\" >Clustering Iteration</td>\n",
       "                        <td id=\"T_50c18_row45_col1\" class=\"data row45 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row46\" class=\"row_heading level0 row46\" >46</th>\n",
       "                        <td id=\"T_50c18_row46_col0\" class=\"data row46 col0\" >Polynomial Features</td>\n",
       "                        <td id=\"T_50c18_row46_col1\" class=\"data row46 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row47\" class=\"row_heading level0 row47\" >47</th>\n",
       "                        <td id=\"T_50c18_row47_col0\" class=\"data row47 col0\" >Polynomial Degree</td>\n",
       "                        <td id=\"T_50c18_row47_col1\" class=\"data row47 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row48\" class=\"row_heading level0 row48\" >48</th>\n",
       "                        <td id=\"T_50c18_row48_col0\" class=\"data row48 col0\" >Trignometry Features</td>\n",
       "                        <td id=\"T_50c18_row48_col1\" class=\"data row48 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row49\" class=\"row_heading level0 row49\" >49</th>\n",
       "                        <td id=\"T_50c18_row49_col0\" class=\"data row49 col0\" >Polynomial Threshold</td>\n",
       "                        <td id=\"T_50c18_row49_col1\" class=\"data row49 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row50\" class=\"row_heading level0 row50\" >50</th>\n",
       "                        <td id=\"T_50c18_row50_col0\" class=\"data row50 col0\" >Group Features</td>\n",
       "                        <td id=\"T_50c18_row50_col1\" class=\"data row50 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row51\" class=\"row_heading level0 row51\" >51</th>\n",
       "                        <td id=\"T_50c18_row51_col0\" class=\"data row51 col0\" >Feature Selection</td>\n",
       "                        <td id=\"T_50c18_row51_col1\" class=\"data row51 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row52\" class=\"row_heading level0 row52\" >52</th>\n",
       "                        <td id=\"T_50c18_row52_col0\" class=\"data row52 col0\" >Feature Selection Method</td>\n",
       "                        <td id=\"T_50c18_row52_col1\" class=\"data row52 col1\" >classic</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row53\" class=\"row_heading level0 row53\" >53</th>\n",
       "                        <td id=\"T_50c18_row53_col0\" class=\"data row53 col0\" >Features Selection Threshold</td>\n",
       "                        <td id=\"T_50c18_row53_col1\" class=\"data row53 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row54\" class=\"row_heading level0 row54\" >54</th>\n",
       "                        <td id=\"T_50c18_row54_col0\" class=\"data row54 col0\" >Feature Interaction</td>\n",
       "                        <td id=\"T_50c18_row54_col1\" class=\"data row54 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row55\" class=\"row_heading level0 row55\" >55</th>\n",
       "                        <td id=\"T_50c18_row55_col0\" class=\"data row55 col0\" >Feature Ratio</td>\n",
       "                        <td id=\"T_50c18_row55_col1\" class=\"data row55 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row56\" class=\"row_heading level0 row56\" >56</th>\n",
       "                        <td id=\"T_50c18_row56_col0\" class=\"data row56 col0\" >Interaction Threshold</td>\n",
       "                        <td id=\"T_50c18_row56_col1\" class=\"data row56 col1\" >None</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row57\" class=\"row_heading level0 row57\" >57</th>\n",
       "                        <td id=\"T_50c18_row57_col0\" class=\"data row57 col0\" >Fix Imbalance</td>\n",
       "                        <td id=\"T_50c18_row57_col1\" class=\"data row57 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_50c18_level0_row58\" class=\"row_heading level0 row58\" >58</th>\n",
       "                        <td id=\"T_50c18_row58_col0\" class=\"data row58 col0\" >Fix Imbalance Method</td>\n",
       "                        <td id=\"T_50c18_row58_col1\" class=\"data row58 col1\" >SMOTE</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7ff2ac5aee90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# import the classification module \n",
    "from pycaret import classification\n",
    "from pycaret.classification import *\n",
    "\n",
    "# setup the environment \n",
    "classification_setup = classification.setup(data = data_classification, target = 'cell_type_idx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "obvious-weekly",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_7906f_row10_col0,#T_7906f_row10_col1,#T_7906f_row10_col2,#T_7906f_row10_col3,#T_7906f_row10_col4,#T_7906f_row10_col5,#T_7906f_row10_col6{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_7906f_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Accuracy</th>        <th class=\"col_heading level0 col1\" >AUC</th>        <th class=\"col_heading level0 col2\" >Recall</th>        <th class=\"col_heading level0 col3\" >Prec.</th>        <th class=\"col_heading level0 col4\" >F1</th>        <th class=\"col_heading level0 col5\" >Kappa</th>        <th class=\"col_heading level0 col6\" >MCC</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_7906f_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_7906f_row0_col0\" class=\"data row0 col0\" >0.7489</td>\n",
       "                        <td id=\"T_7906f_row0_col1\" class=\"data row0 col1\" >0.9283</td>\n",
       "                        <td id=\"T_7906f_row0_col2\" class=\"data row0 col2\" >0.4158</td>\n",
       "                        <td id=\"T_7906f_row0_col3\" class=\"data row0 col3\" >0.7296</td>\n",
       "                        <td id=\"T_7906f_row0_col4\" class=\"data row0 col4\" >0.7363</td>\n",
       "                        <td id=\"T_7906f_row0_col5\" class=\"data row0 col5\" >0.4957</td>\n",
       "                        <td id=\"T_7906f_row0_col6\" class=\"data row0 col6\" >0.4977</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_7906f_row1_col0\" class=\"data row1 col0\" >0.7805</td>\n",
       "                        <td id=\"T_7906f_row1_col1\" class=\"data row1 col1\" >0.9356</td>\n",
       "                        <td id=\"T_7906f_row1_col2\" class=\"data row1 col2\" >0.4646</td>\n",
       "                        <td id=\"T_7906f_row1_col3\" class=\"data row1 col3\" >0.7656</td>\n",
       "                        <td id=\"T_7906f_row1_col4\" class=\"data row1 col4\" >0.7658</td>\n",
       "                        <td id=\"T_7906f_row1_col5\" class=\"data row1 col5\" >0.5488</td>\n",
       "                        <td id=\"T_7906f_row1_col6\" class=\"data row1 col6\" >0.5533</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_7906f_row2_col0\" class=\"data row2 col0\" >0.7747</td>\n",
       "                        <td id=\"T_7906f_row2_col1\" class=\"data row2 col1\" >0.9292</td>\n",
       "                        <td id=\"T_7906f_row2_col2\" class=\"data row2 col2\" >0.4742</td>\n",
       "                        <td id=\"T_7906f_row2_col3\" class=\"data row2 col3\" >0.7550</td>\n",
       "                        <td id=\"T_7906f_row2_col4\" class=\"data row2 col4\" >0.7609</td>\n",
       "                        <td id=\"T_7906f_row2_col5\" class=\"data row2 col5\" >0.5485</td>\n",
       "                        <td id=\"T_7906f_row2_col6\" class=\"data row2 col6\" >0.5511</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_7906f_row3_col0\" class=\"data row3 col0\" >0.7719</td>\n",
       "                        <td id=\"T_7906f_row3_col1\" class=\"data row3 col1\" >0.9384</td>\n",
       "                        <td id=\"T_7906f_row3_col2\" class=\"data row3 col2\" >0.4592</td>\n",
       "                        <td id=\"T_7906f_row3_col3\" class=\"data row3 col3\" >0.7543</td>\n",
       "                        <td id=\"T_7906f_row3_col4\" class=\"data row3 col4\" >0.7564</td>\n",
       "                        <td id=\"T_7906f_row3_col5\" class=\"data row3 col5\" >0.5295</td>\n",
       "                        <td id=\"T_7906f_row3_col6\" class=\"data row3 col6\" >0.5342</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_7906f_row4_col0\" class=\"data row4 col0\" >0.7647</td>\n",
       "                        <td id=\"T_7906f_row4_col1\" class=\"data row4 col1\" >0.9414</td>\n",
       "                        <td id=\"T_7906f_row4_col2\" class=\"data row4 col2\" >0.4711</td>\n",
       "                        <td id=\"T_7906f_row4_col3\" class=\"data row4 col3\" >0.7429</td>\n",
       "                        <td id=\"T_7906f_row4_col4\" class=\"data row4 col4\" >0.7474</td>\n",
       "                        <td id=\"T_7906f_row4_col5\" class=\"data row4 col5\" >0.5110</td>\n",
       "                        <td id=\"T_7906f_row4_col6\" class=\"data row4 col6\" >0.5170</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row5\" class=\"row_heading level0 row5\" >5</th>\n",
       "                        <td id=\"T_7906f_row5_col0\" class=\"data row5 col0\" >0.7547</td>\n",
       "                        <td id=\"T_7906f_row5_col1\" class=\"data row5 col1\" >0.9367</td>\n",
       "                        <td id=\"T_7906f_row5_col2\" class=\"data row5 col2\" >0.4330</td>\n",
       "                        <td id=\"T_7906f_row5_col3\" class=\"data row5 col3\" >0.7301</td>\n",
       "                        <td id=\"T_7906f_row5_col4\" class=\"data row5 col4\" >0.7385</td>\n",
       "                        <td id=\"T_7906f_row5_col5\" class=\"data row5 col5\" >0.4963</td>\n",
       "                        <td id=\"T_7906f_row5_col6\" class=\"data row5 col6\" >0.5004</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row6\" class=\"row_heading level0 row6\" >6</th>\n",
       "                        <td id=\"T_7906f_row6_col0\" class=\"data row6 col0\" >0.7504</td>\n",
       "                        <td id=\"T_7906f_row6_col1\" class=\"data row6 col1\" >0.9244</td>\n",
       "                        <td id=\"T_7906f_row6_col2\" class=\"data row6 col2\" >0.4364</td>\n",
       "                        <td id=\"T_7906f_row6_col3\" class=\"data row6 col3\" >0.7311</td>\n",
       "                        <td id=\"T_7906f_row6_col4\" class=\"data row6 col4\" >0.7360</td>\n",
       "                        <td id=\"T_7906f_row6_col5\" class=\"data row6 col5\" >0.4885</td>\n",
       "                        <td id=\"T_7906f_row6_col6\" class=\"data row6 col6\" >0.4923</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row7\" class=\"row_heading level0 row7\" >7</th>\n",
       "                        <td id=\"T_7906f_row7_col0\" class=\"data row7 col0\" >0.7661</td>\n",
       "                        <td id=\"T_7906f_row7_col1\" class=\"data row7 col1\" >0.9263</td>\n",
       "                        <td id=\"T_7906f_row7_col2\" class=\"data row7 col2\" >0.4197</td>\n",
       "                        <td id=\"T_7906f_row7_col3\" class=\"data row7 col3\" >0.7417</td>\n",
       "                        <td id=\"T_7906f_row7_col4\" class=\"data row7 col4\" >0.7469</td>\n",
       "                        <td id=\"T_7906f_row7_col5\" class=\"data row7 col5\" >0.5133</td>\n",
       "                        <td id=\"T_7906f_row7_col6\" class=\"data row7 col6\" >0.5199</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row8\" class=\"row_heading level0 row8\" >8</th>\n",
       "                        <td id=\"T_7906f_row8_col0\" class=\"data row8 col0\" >0.7862</td>\n",
       "                        <td id=\"T_7906f_row8_col1\" class=\"data row8 col1\" >0.9397</td>\n",
       "                        <td id=\"T_7906f_row8_col2\" class=\"data row8 col2\" >0.4857</td>\n",
       "                        <td id=\"T_7906f_row8_col3\" class=\"data row8 col3\" >0.7604</td>\n",
       "                        <td id=\"T_7906f_row8_col4\" class=\"data row8 col4\" >0.7682</td>\n",
       "                        <td id=\"T_7906f_row8_col5\" class=\"data row8 col5\" >0.5552</td>\n",
       "                        <td id=\"T_7906f_row8_col6\" class=\"data row8 col6\" >0.5622</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row9\" class=\"row_heading level0 row9\" >9</th>\n",
       "                        <td id=\"T_7906f_row9_col0\" class=\"data row9 col0\" >0.7618</td>\n",
       "                        <td id=\"T_7906f_row9_col1\" class=\"data row9 col1\" >0.9323</td>\n",
       "                        <td id=\"T_7906f_row9_col2\" class=\"data row9 col2\" >0.4591</td>\n",
       "                        <td id=\"T_7906f_row9_col3\" class=\"data row9 col3\" >0.7398</td>\n",
       "                        <td id=\"T_7906f_row9_col4\" class=\"data row9 col4\" >0.7425</td>\n",
       "                        <td id=\"T_7906f_row9_col5\" class=\"data row9 col5\" >0.5011</td>\n",
       "                        <td id=\"T_7906f_row9_col6\" class=\"data row9 col6\" >0.5086</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row10\" class=\"row_heading level0 row10\" >Mean</th>\n",
       "                        <td id=\"T_7906f_row10_col0\" class=\"data row10 col0\" >0.7660</td>\n",
       "                        <td id=\"T_7906f_row10_col1\" class=\"data row10 col1\" >0.9332</td>\n",
       "                        <td id=\"T_7906f_row10_col2\" class=\"data row10 col2\" >0.4519</td>\n",
       "                        <td id=\"T_7906f_row10_col3\" class=\"data row10 col3\" >0.7450</td>\n",
       "                        <td id=\"T_7906f_row10_col4\" class=\"data row10 col4\" >0.7499</td>\n",
       "                        <td id=\"T_7906f_row10_col5\" class=\"data row10 col5\" >0.5188</td>\n",
       "                        <td id=\"T_7906f_row10_col6\" class=\"data row10 col6\" >0.5237</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7906f_level0_row11\" class=\"row_heading level0 row11\" >SD</th>\n",
       "                        <td id=\"T_7906f_row11_col0\" class=\"data row11 col0\" >0.0119</td>\n",
       "                        <td id=\"T_7906f_row11_col1\" class=\"data row11 col1\" >0.0057</td>\n",
       "                        <td id=\"T_7906f_row11_col2\" class=\"data row11 col2\" >0.0228</td>\n",
       "                        <td id=\"T_7906f_row11_col3\" class=\"data row11 col3\" >0.0124</td>\n",
       "                        <td id=\"T_7906f_row11_col4\" class=\"data row11 col4\" >0.0115</td>\n",
       "                        <td id=\"T_7906f_row11_col5\" class=\"data row11 col5\" >0.0236</td>\n",
       "                        <td id=\"T_7906f_row11_col6\" class=\"data row11 col6\" >0.0239</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7ff2ab5ae8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# build the xgboost model\n",
    "\n",
    "classification_xgb = classification.create_model('xgboost')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "retired-recruitment",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_labels, y_label, train_size = 0.8, random_state = 40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "painful-external",
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "social-detroit",
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier = xgb.XGBClassifier(base_score=0.5, booster='gbtree',\n",
    "                                            colsample_bylevel=1,\n",
    "                                            colsample_bynode=1,\n",
    "                                            colsample_bytree=1, gamma=0,\n",
    "                                            learning_rate=0.1, max_delta_step=0,\n",
    "                                            max_depth=3, min_child_weight=1,\n",
    "                                            missing=None, n_estimators=100,\n",
    "                                            n_jobs=-1, nthread=None,\n",
    "                                            objective='binary:logistic',\n",
    "                                            random_state=1855, reg_alpha=0,\n",
    "                                            reg_lambda=1, scale_pos_weight=1,\n",
    "                                            seed=None, silent=None, subsample=1,\n",
    "                                            verbosity=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "multiple-behalf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,\n",
       "              importance_type='gain', interaction_constraints='',\n",
       "              learning_rate=0.1, max_delta_step=0, max_depth=3,\n",
       "              min_child_weight=1, missing=None, monotone_constraints='()',\n",
       "              n_estimators=100, n_jobs=-1, nthread=4, num_parallel_tree=1,\n",
       "              objective='multi:softprob', random_state=1855, reg_alpha=0,\n",
       "              reg_lambda=1, scale_pos_weight=1, seed=1855, silent=None,\n",
       "              subsample=1, tree_method='exact', use_label_encoder=True,\n",
       "              validate_parameters=1, ...)"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classifier.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "exact-legislation",
   "metadata": {},
   "source": [
    "## CNN + Xgboost evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "interested-making",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = classifier.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "joint-protocol",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  21,    9,   20,    1,   16,    9,    0],\n",
       "       [   5,   40,    6,    0,   35,   10,    0],\n",
       "       [   7,   14,   96,    0,   74,   29,    1],\n",
       "       [   0,    0,    2,    8,   18,    0,    0],\n",
       "       [   2,    7,   24,    0, 1271,   24,    0],\n",
       "       [   4,    7,   14,    0,  104,   86,    1],\n",
       "       [   1,    1,    1,    0,   14,    1,    9]])"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "cubic-speaking",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.53      0.28      0.36        76\n",
      "           1       0.51      0.42      0.46        96\n",
      "           2       0.59      0.43      0.50       221\n",
      "           3       0.89      0.29      0.43        28\n",
      "           4       0.83      0.96      0.89      1328\n",
      "           5       0.54      0.40      0.46       216\n",
      "           6       0.82      0.33      0.47        27\n",
      "\n",
      "    accuracy                           0.77      1992\n",
      "   macro avg       0.67      0.44      0.51      1992\n",
      "weighted avg       0.75      0.77      0.75      1992\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "id": "secondary-twist",
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier.save_model('classifier_v1.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "dominant-flour",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "file_name = \"xgb_classifier.pkl\"\n",
    "\n",
    "# save\n",
    "pickle.dump(classifier, open(file_name, \"wb\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "classified-boost",
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}