Epileptic Seizure Classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **Epileptic Seizure Classification**\n",
    "\n",
    "**Dataset:** Epileptic Seizure Recognition dataset available at https://archive.ics.uci.edu/ml/datasets/Epileptic+Seizure+Recognition\n",
    "\n",
    "**Code:** Chapter 3 Classification from `Hands-On Machine Learning with Scikit-Learn and TensorFlow` book (Aurélien Géron)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Common imports\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get the data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predict the `y` class = {1, 2, 3, 4, 5} "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dataset attributes as described on UCI:\n",
    "The response variable is y in column 179, the Explanatory variables X1, X2, ..., X178\n",
    "\n",
    "y contains the category of the 178-dimensional input vector. Specifically y in {1, 2, 3, 4, 5}:\n",
    "\n",
    "5 - eyes open, means when they were recording the EEG signal of the brain the patient had their eyes open\n",
    "\n",
    "4 - eyes closed, means when they were recording the EEG signal the patient had their eyes closed\n",
    "\n",
    "3 - Yes they identify where the region of the tumor was in the brain and recording the EEG activity from the healthy brain area\n",
    "\n",
    "2 - They recorder the EEG from the area where the tumor was located\n",
    "\n",
    "1 - Recording of seizure activity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "EPILEPTIC_SEIZURE_PATH = \"dataset\"\n",
    "\n",
    "def load_epileptic_seizure_data(epileptic_seizure_path=EPILEPTIC_SEIZURE_PATH):\n",
    "    csv_path = os.path.join(epileptic_seizure_path, \"data.csv\")\n",
    "    return pd.read_csv(csv_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>X1</th>\n",
       "      <th>X2</th>\n",
       "      <th>X3</th>\n",
       "      <th>X4</th>\n",
       "      <th>X5</th>\n",
       "      <th>X6</th>\n",
       "      <th>X7</th>\n",
       "      <th>X8</th>\n",
       "      <th>X9</th>\n",
       "      <th>...</th>\n",
       "      <th>X170</th>\n",
       "      <th>X171</th>\n",
       "      <th>X172</th>\n",
       "      <th>X173</th>\n",
       "      <th>X174</th>\n",
       "      <th>X175</th>\n",
       "      <th>X176</th>\n",
       "      <th>X177</th>\n",
       "      <th>X178</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>X21.V1.791</td>\n",
       "      <td>135</td>\n",
       "      <td>190</td>\n",
       "      <td>229</td>\n",
       "      <td>223</td>\n",
       "      <td>192</td>\n",
       "      <td>125</td>\n",
       "      <td>55</td>\n",
       "      <td>-9</td>\n",
       "      <td>-33</td>\n",
       "      <td>...</td>\n",
       "      <td>-17</td>\n",
       "      <td>-15</td>\n",
       "      <td>-31</td>\n",
       "      <td>-77</td>\n",
       "      <td>-103</td>\n",
       "      <td>-127</td>\n",
       "      <td>-116</td>\n",
       "      <td>-83</td>\n",
       "      <td>-51</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>X15.V1.924</td>\n",
       "      <td>386</td>\n",
       "      <td>382</td>\n",
       "      <td>356</td>\n",
       "      <td>331</td>\n",
       "      <td>320</td>\n",
       "      <td>315</td>\n",
       "      <td>307</td>\n",
       "      <td>272</td>\n",
       "      <td>244</td>\n",
       "      <td>...</td>\n",
       "      <td>164</td>\n",
       "      <td>150</td>\n",
       "      <td>146</td>\n",
       "      <td>152</td>\n",
       "      <td>157</td>\n",
       "      <td>156</td>\n",
       "      <td>154</td>\n",
       "      <td>143</td>\n",
       "      <td>129</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>X8.V1.1</td>\n",
       "      <td>-32</td>\n",
       "      <td>-39</td>\n",
       "      <td>-47</td>\n",
       "      <td>-37</td>\n",
       "      <td>-32</td>\n",
       "      <td>-36</td>\n",
       "      <td>-57</td>\n",
       "      <td>-73</td>\n",
       "      <td>-85</td>\n",
       "      <td>...</td>\n",
       "      <td>57</td>\n",
       "      <td>64</td>\n",
       "      <td>48</td>\n",
       "      <td>19</td>\n",
       "      <td>-12</td>\n",
       "      <td>-30</td>\n",
       "      <td>-35</td>\n",
       "      <td>-35</td>\n",
       "      <td>-36</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>X16.V1.60</td>\n",
       "      <td>-105</td>\n",
       "      <td>-101</td>\n",
       "      <td>-96</td>\n",
       "      <td>-92</td>\n",
       "      <td>-89</td>\n",
       "      <td>-95</td>\n",
       "      <td>-102</td>\n",
       "      <td>-100</td>\n",
       "      <td>-87</td>\n",
       "      <td>...</td>\n",
       "      <td>-82</td>\n",
       "      <td>-81</td>\n",
       "      <td>-80</td>\n",
       "      <td>-77</td>\n",
       "      <td>-85</td>\n",
       "      <td>-77</td>\n",
       "      <td>-72</td>\n",
       "      <td>-69</td>\n",
       "      <td>-65</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>X20.V1.54</td>\n",
       "      <td>-9</td>\n",
       "      <td>-65</td>\n",
       "      <td>-98</td>\n",
       "      <td>-102</td>\n",
       "      <td>-78</td>\n",
       "      <td>-48</td>\n",
       "      <td>-16</td>\n",
       "      <td>0</td>\n",
       "      <td>-21</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>-12</td>\n",
       "      <td>-32</td>\n",
       "      <td>-41</td>\n",
       "      <td>-65</td>\n",
       "      <td>-83</td>\n",
       "      <td>-89</td>\n",
       "      <td>-73</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 180 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0   X1   X2   X3   X4   X5   X6   X7   X8   X9  ...  X170  X171  \\\n",
       "0  X21.V1.791  135  190  229  223  192  125   55   -9  -33  ...   -17   -15   \n",
       "1  X15.V1.924  386  382  356  331  320  315  307  272  244  ...   164   150   \n",
       "2     X8.V1.1  -32  -39  -47  -37  -32  -36  -57  -73  -85  ...    57    64   \n",
       "3   X16.V1.60 -105 -101  -96  -92  -89  -95 -102 -100  -87  ...   -82   -81   \n",
       "4   X20.V1.54   -9  -65  -98 -102  -78  -48  -16    0  -21  ...     4     2   \n",
       "\n",
       "   X172  X173  X174  X175  X176  X177  X178  y  \n",
       "0   -31   -77  -103  -127  -116   -83   -51  4  \n",
       "1   146   152   157   156   154   143   129  1  \n",
       "2    48    19   -12   -30   -35   -35   -36  5  \n",
       "3   -80   -77   -85   -77   -72   -69   -65  5  \n",
       "4   -12   -32   -41   -65   -83   -89   -73  5  \n",
       "\n",
       "[5 rows x 180 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "epileptic_seizure = load_epileptic_seizure_data()\n",
    "epileptic_seizure.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Give a name to the first column, as it is unnamed:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "epileptic_seizure.rename( columns={'Unnamed: 0':'ID'}, inplace=True )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 11500 entries, 0 to 11499\n",
      "Columns: 180 entries, ID to y\n",
      "dtypes: int64(179), object(1)\n",
      "memory usage: 15.8+ MB\n"
     ]
    }
   ],
   "source": [
    "epileptic_seizure.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5    2300\n",
       "4    2300\n",
       "3    2300\n",
       "2    2300\n",
       "1    2300\n",
       "Name: y, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "epileptic_seizure[\"y\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>X1</th>\n",
       "      <th>X2</th>\n",
       "      <th>X3</th>\n",
       "      <th>X4</th>\n",
       "      <th>X5</th>\n",
       "      <th>X6</th>\n",
       "      <th>X7</th>\n",
       "      <th>X8</th>\n",
       "      <th>X9</th>\n",
       "      <th>X10</th>\n",
       "      <th>...</th>\n",
       "      <th>X170</th>\n",
       "      <th>X171</th>\n",
       "      <th>X172</th>\n",
       "      <th>X173</th>\n",
       "      <th>X174</th>\n",
       "      <th>X175</th>\n",
       "      <th>X176</th>\n",
       "      <th>X177</th>\n",
       "      <th>X178</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.00000</td>\n",
       "      <td>11500.00000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "      <td>11500.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>-11.581391</td>\n",
       "      <td>-10.911565</td>\n",
       "      <td>-10.187130</td>\n",
       "      <td>-9.143043</td>\n",
       "      <td>-8.009739</td>\n",
       "      <td>-7.003478</td>\n",
       "      <td>-6.502087</td>\n",
       "      <td>-6.68713</td>\n",
       "      <td>-6.55800</td>\n",
       "      <td>-6.168435</td>\n",
       "      <td>...</td>\n",
       "      <td>-10.145739</td>\n",
       "      <td>-11.630348</td>\n",
       "      <td>-12.943478</td>\n",
       "      <td>-13.668870</td>\n",
       "      <td>-13.363304</td>\n",
       "      <td>-13.045043</td>\n",
       "      <td>-12.705130</td>\n",
       "      <td>-12.426000</td>\n",
       "      <td>-12.195652</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>165.626284</td>\n",
       "      <td>166.059609</td>\n",
       "      <td>163.524317</td>\n",
       "      <td>161.269041</td>\n",
       "      <td>160.998007</td>\n",
       "      <td>161.328725</td>\n",
       "      <td>161.467837</td>\n",
       "      <td>162.11912</td>\n",
       "      <td>162.03336</td>\n",
       "      <td>160.436352</td>\n",
       "      <td>...</td>\n",
       "      <td>164.652883</td>\n",
       "      <td>166.149790</td>\n",
       "      <td>168.554058</td>\n",
       "      <td>168.556486</td>\n",
       "      <td>167.257290</td>\n",
       "      <td>164.241019</td>\n",
       "      <td>162.895832</td>\n",
       "      <td>162.886311</td>\n",
       "      <td>164.852015</td>\n",
       "      <td>1.414275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-1839.000000</td>\n",
       "      <td>-1838.000000</td>\n",
       "      <td>-1835.000000</td>\n",
       "      <td>-1845.000000</td>\n",
       "      <td>-1791.000000</td>\n",
       "      <td>-1757.000000</td>\n",
       "      <td>-1832.000000</td>\n",
       "      <td>-1778.00000</td>\n",
       "      <td>-1840.00000</td>\n",
       "      <td>-1867.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-1867.000000</td>\n",
       "      <td>-1865.000000</td>\n",
       "      <td>-1642.000000</td>\n",
       "      <td>-1723.000000</td>\n",
       "      <td>-1866.000000</td>\n",
       "      <td>-1863.000000</td>\n",
       "      <td>-1781.000000</td>\n",
       "      <td>-1727.000000</td>\n",
       "      <td>-1829.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-54.000000</td>\n",
       "      <td>-55.000000</td>\n",
       "      <td>-54.000000</td>\n",
       "      <td>-54.000000</td>\n",
       "      <td>-54.000000</td>\n",
       "      <td>-54.000000</td>\n",
       "      <td>-54.000000</td>\n",
       "      <td>-55.00000</td>\n",
       "      <td>-55.00000</td>\n",
       "      <td>-54.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-55.000000</td>\n",
       "      <td>-56.000000</td>\n",
       "      <td>-56.000000</td>\n",
       "      <td>-56.000000</td>\n",
       "      <td>-55.000000</td>\n",
       "      <td>-56.000000</td>\n",
       "      <td>-55.000000</td>\n",
       "      <td>-55.000000</td>\n",
       "      <td>-55.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-8.000000</td>\n",
       "      <td>-8.000000</td>\n",
       "      <td>-7.000000</td>\n",
       "      <td>-8.000000</td>\n",
       "      <td>-8.000000</td>\n",
       "      <td>-8.000000</td>\n",
       "      <td>-8.000000</td>\n",
       "      <td>-8.00000</td>\n",
       "      <td>-7.00000</td>\n",
       "      <td>-7.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-9.000000</td>\n",
       "      <td>-10.000000</td>\n",
       "      <td>-10.000000</td>\n",
       "      <td>-10.000000</td>\n",
       "      <td>-10.000000</td>\n",
       "      <td>-9.000000</td>\n",
       "      <td>-9.000000</td>\n",
       "      <td>-9.000000</td>\n",
       "      <td>-9.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>34.000000</td>\n",
       "      <td>35.000000</td>\n",
       "      <td>36.000000</td>\n",
       "      <td>36.000000</td>\n",
       "      <td>35.000000</td>\n",
       "      <td>36.000000</td>\n",
       "      <td>35.000000</td>\n",
       "      <td>36.00000</td>\n",
       "      <td>36.00000</td>\n",
       "      <td>35.250000</td>\n",
       "      <td>...</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1726.000000</td>\n",
       "      <td>1713.000000</td>\n",
       "      <td>1697.000000</td>\n",
       "      <td>1612.000000</td>\n",
       "      <td>1518.000000</td>\n",
       "      <td>1816.000000</td>\n",
       "      <td>2047.000000</td>\n",
       "      <td>2047.00000</td>\n",
       "      <td>2047.00000</td>\n",
       "      <td>2047.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1777.000000</td>\n",
       "      <td>1472.000000</td>\n",
       "      <td>1319.000000</td>\n",
       "      <td>1436.000000</td>\n",
       "      <td>1733.000000</td>\n",
       "      <td>1958.000000</td>\n",
       "      <td>2047.000000</td>\n",
       "      <td>2047.000000</td>\n",
       "      <td>1915.000000</td>\n",
       "      <td>5.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 179 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 X1            X2            X3            X4            X5  \\\n",
       "count  11500.000000  11500.000000  11500.000000  11500.000000  11500.000000   \n",
       "mean     -11.581391    -10.911565    -10.187130     -9.143043     -8.009739   \n",
       "std      165.626284    166.059609    163.524317    161.269041    160.998007   \n",
       "min    -1839.000000  -1838.000000  -1835.000000  -1845.000000  -1791.000000   \n",
       "25%      -54.000000    -55.000000    -54.000000    -54.000000    -54.000000   \n",
       "50%       -8.000000     -8.000000     -7.000000     -8.000000     -8.000000   \n",
       "75%       34.000000     35.000000     36.000000     36.000000     35.000000   \n",
       "max     1726.000000   1713.000000   1697.000000   1612.000000   1518.000000   \n",
       "\n",
       "                 X6            X7           X8           X9           X10  \\\n",
       "count  11500.000000  11500.000000  11500.00000  11500.00000  11500.000000   \n",
       "mean      -7.003478     -6.502087     -6.68713     -6.55800     -6.168435   \n",
       "std      161.328725    161.467837    162.11912    162.03336    160.436352   \n",
       "min    -1757.000000  -1832.000000  -1778.00000  -1840.00000  -1867.000000   \n",
       "25%      -54.000000    -54.000000    -55.00000    -55.00000    -54.000000   \n",
       "50%       -8.000000     -8.000000     -8.00000     -7.00000     -7.000000   \n",
       "75%       36.000000     35.000000     36.00000     36.00000     35.250000   \n",
       "max     1816.000000   2047.000000   2047.00000   2047.00000   2047.000000   \n",
       "\n",
       "       ...          X170          X171          X172          X173  \\\n",
       "count  ...  11500.000000  11500.000000  11500.000000  11500.000000   \n",
       "mean   ...    -10.145739    -11.630348    -12.943478    -13.668870   \n",
       "std    ...    164.652883    166.149790    168.554058    168.556486   \n",
       "min    ...  -1867.000000  -1865.000000  -1642.000000  -1723.000000   \n",
       "25%    ...    -55.000000    -56.000000    -56.000000    -56.000000   \n",
       "50%    ...     -9.000000    -10.000000    -10.000000    -10.000000   \n",
       "75%    ...     34.000000     34.000000     33.000000     33.000000   \n",
       "max    ...   1777.000000   1472.000000   1319.000000   1436.000000   \n",
       "\n",
       "               X174          X175          X176          X177          X178  \\\n",
       "count  11500.000000  11500.000000  11500.000000  11500.000000  11500.000000   \n",
       "mean     -13.363304    -13.045043    -12.705130    -12.426000    -12.195652   \n",
       "std      167.257290    164.241019    162.895832    162.886311    164.852015   \n",
       "min    -1866.000000  -1863.000000  -1781.000000  -1727.000000  -1829.000000   \n",
       "25%      -55.000000    -56.000000    -55.000000    -55.000000    -55.000000   \n",
       "50%      -10.000000     -9.000000     -9.000000     -9.000000     -9.000000   \n",
       "75%       34.000000     34.000000     34.000000     34.000000     34.000000   \n",
       "max     1733.000000   1958.000000   2047.000000   2047.000000   1915.000000   \n",
       "\n",
       "                  y  \n",
       "count  11500.000000  \n",
       "mean       3.000000  \n",
       "std        1.414275  \n",
       "min        1.000000  \n",
       "25%        2.000000  \n",
       "50%        3.000000  \n",
       "75%        4.000000  \n",
       "max        5.000000  \n",
       "\n",
       "[8 rows x 179 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "epileptic_seizure.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x15395091248>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "epileptic_seizure[\"X55\"].hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "epileptic_seizure.plot(kind=\"scatter\", x=\"X55\", y=\"y\", figsize=(10,7), alpha=0.1)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop the `ID` attribute, as it is the ID of the record:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "epileptic_seizure = epileptic_seizure.drop(\"ID\", axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr_matrix = epileptic_seizure.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "y       1.000000\n",
       "X173    0.029172\n",
       "X174    0.028563\n",
       "X175    0.028425\n",
       "X176    0.027259\n",
       "          ...   \n",
       "X158   -0.034805\n",
       "X12    -0.035379\n",
       "X11    -0.038056\n",
       "X43    -0.038320\n",
       "X44    -0.040499\n",
       "Name: y, Length: 179, dtype: float64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_matrix[\"y\"].sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**No correlation !**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Split the data in training and test sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "train_set, test_set = train_test_split(epileptic_seizure, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>X1</th>\n",
       "      <th>X2</th>\n",
       "      <th>X3</th>\n",
       "      <th>X4</th>\n",
       "      <th>X5</th>\n",
       "      <th>X6</th>\n",
       "      <th>X7</th>\n",
       "      <th>X8</th>\n",
       "      <th>X9</th>\n",
       "      <th>X10</th>\n",
       "      <th>...</th>\n",
       "      <th>X170</th>\n",
       "      <th>X171</th>\n",
       "      <th>X172</th>\n",
       "      <th>X173</th>\n",
       "      <th>X174</th>\n",
       "      <th>X175</th>\n",
       "      <th>X176</th>\n",
       "      <th>X177</th>\n",
       "      <th>X178</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5556</th>\n",
       "      <td>-5</td>\n",
       "      <td>-13</td>\n",
       "      <td>-4</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>12</td>\n",
       "      <td>28</td>\n",
       "      <td>45</td>\n",
       "      <td>58</td>\n",
       "      <td>...</td>\n",
       "      <td>-54</td>\n",
       "      <td>-64</td>\n",
       "      <td>-49</td>\n",
       "      <td>-34</td>\n",
       "      <td>-17</td>\n",
       "      <td>-18</td>\n",
       "      <td>-10</td>\n",
       "      <td>-10</td>\n",
       "      <td>-15</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1570</th>\n",
       "      <td>-39</td>\n",
       "      <td>-32</td>\n",
       "      <td>-21</td>\n",
       "      <td>-14</td>\n",
       "      <td>-9</td>\n",
       "      <td>8</td>\n",
       "      <td>38</td>\n",
       "      <td>67</td>\n",
       "      <td>89</td>\n",
       "      <td>87</td>\n",
       "      <td>...</td>\n",
       "      <td>-14</td>\n",
       "      <td>-71</td>\n",
       "      <td>-123</td>\n",
       "      <td>-143</td>\n",
       "      <td>-154</td>\n",
       "      <td>-129</td>\n",
       "      <td>-94</td>\n",
       "      <td>-46</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2582</th>\n",
       "      <td>115</td>\n",
       "      <td>106</td>\n",
       "      <td>39</td>\n",
       "      <td>-4</td>\n",
       "      <td>-10</td>\n",
       "      <td>-10</td>\n",
       "      <td>-22</td>\n",
       "      <td>-20</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>-55</td>\n",
       "      <td>-29</td>\n",
       "      <td>-14</td>\n",
       "      <td>-18</td>\n",
       "      <td>-29</td>\n",
       "      <td>-29</td>\n",
       "      <td>-15</td>\n",
       "      <td>-20</td>\n",
       "      <td>-10</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4948</th>\n",
       "      <td>-6</td>\n",
       "      <td>-14</td>\n",
       "      <td>-25</td>\n",
       "      <td>-24</td>\n",
       "      <td>-10</td>\n",
       "      <td>9</td>\n",
       "      <td>26</td>\n",
       "      <td>54</td>\n",
       "      <td>61</td>\n",
       "      <td>64</td>\n",
       "      <td>...</td>\n",
       "      <td>-7</td>\n",
       "      <td>-24</td>\n",
       "      <td>-9</td>\n",
       "      <td>32</td>\n",
       "      <td>65</td>\n",
       "      <td>97</td>\n",
       "      <td>100</td>\n",
       "      <td>88</td>\n",
       "      <td>80</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7089</th>\n",
       "      <td>59</td>\n",
       "      <td>56</td>\n",
       "      <td>71</td>\n",
       "      <td>58</td>\n",
       "      <td>66</td>\n",
       "      <td>52</td>\n",
       "      <td>34</td>\n",
       "      <td>17</td>\n",
       "      <td>-2</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>-29</td>\n",
       "      <td>3</td>\n",
       "      <td>38</td>\n",
       "      <td>62</td>\n",
       "      <td>65</td>\n",
       "      <td>40</td>\n",
       "      <td>18</td>\n",
       "      <td>5</td>\n",
       "      <td>-6</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 179 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       X1   X2  X3  X4  X5  X6  X7  X8  X9  X10  ...  X170  X171  X172  X173  \\\n",
       "5556   -5  -13  -4   3   4   4  12  28  45   58  ...   -54   -64   -49   -34   \n",
       "1570  -39  -32 -21 -14  -9   8  38  67  89   87  ...   -14   -71  -123  -143   \n",
       "2582  115  106  39  -4 -10 -10 -22 -20   1    6  ...   -55   -29   -14   -18   \n",
       "4948   -6  -14 -25 -24 -10   9  26  54  61   64  ...    -7   -24    -9    32   \n",
       "7089   59   56  71  58  66  52  34  17  -2    0  ...   -29     3    38    62   \n",
       "\n",
       "      X174  X175  X176  X177  X178  y  \n",
       "5556   -17   -18   -10   -10   -15  5  \n",
       "1570  -154  -129   -94   -46     1  4  \n",
       "2582   -29   -29   -15   -20   -10  4  \n",
       "4948    65    97   100    88    80  5  \n",
       "7089    65    40    18     5    -6  5  \n",
       "\n",
       "[5 rows x 179 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_set.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = train_set.drop(\"y\", axis=1) # drop labels for training set\n",
    "y_train = train_set[\"y\"].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = test_set.drop(\"y\", axis=1) # drop labels for training set\n",
    "y_test = test_set[\"y\"].copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Shuffle the training data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "shuffle_index = np.random.permutation(X_train.shape[0])\n",
    "X_train, y_train = X_train.iloc[shuffle_index], y_train.iloc[shuffle_index]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Expecting no incomplete rows, and **no imputation** later in the pipeline:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>X1</th>\n",
       "      <th>X2</th>\n",
       "      <th>X3</th>\n",
       "      <th>X4</th>\n",
       "      <th>X5</th>\n",
       "      <th>X6</th>\n",
       "      <th>X7</th>\n",
       "      <th>X8</th>\n",
       "      <th>X9</th>\n",
       "      <th>X10</th>\n",
       "      <th>...</th>\n",
       "      <th>X169</th>\n",
       "      <th>X170</th>\n",
       "      <th>X171</th>\n",
       "      <th>X172</th>\n",
       "      <th>X173</th>\n",
       "      <th>X174</th>\n",
       "      <th>X175</th>\n",
       "      <th>X176</th>\n",
       "      <th>X177</th>\n",
       "      <th>X178</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>0 rows × 178 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X12, X13, X14, X15, X16, X17, X18, X19, X20, X21, X22, X23, X24, X25, X26, X27, X28, X29, X30, X31, X32, X33, X34, X35, X36, X37, X38, X39, X40, X41, X42, X43, X44, X45, X46, X47, X48, X49, X50, X51, X52, X53, X54, X55, X56, X57, X58, X59, X60, X61, X62, X63, X64, X65, X66, X67, X68, X69, X70, X71, X72, X73, X74, X75, X76, X77, X78, X79, X80, X81, X82, X83, X84, X85, X86, X87, X88, X89, X90, X91, X92, X93, X94, X95, X96, X97, X98, X99, X100, ...]\n",
       "Index: []\n",
       "\n",
       "[0 rows x 178 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_incomplete_rows = X_train[X_train.isnull().any(axis=1)].head()\n",
    "sample_incomplete_rows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "num_pipeline = Pipeline([\n",
    "        ('std_scaler', StandardScaler()),\n",
    "    ])\n",
    "\n",
    "X_train_prepared = num_pipeline.fit_transform(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-0.06680049,  0.03184878,  0.22650231, ...,  0.4175809 ,\n",
       "         0.4567377 ,  0.53786168],\n",
       "       [ 0.06094484,  0.01355136,  0.0099789 , ...,  0.16183686,\n",
       "         0.12030471,  0.06624415],\n",
       "       [-0.10938227, -0.11453057, -0.03951216, ...,  0.02178656,\n",
       "        -0.10602293, -0.19712667],\n",
       "       ...,\n",
       "       [-0.90626983, -1.34045762, -1.344839  , ..., -0.18524433,\n",
       "        -0.17942649,  0.1581177 ],\n",
       "       [ 0.57800929,  0.59296962,  0.61624445, ..., -0.50796894,\n",
       "        -0.41798806, -0.39924847],\n",
       "       [-0.7785245 , -0.72444453, -0.66433685, ..., -0.13653118,\n",
       "        -0.09990597, -0.09912822]])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_prepared"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9200, 178)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_prepared.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Binary Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The seizure is represented by the class 1. Start with a binary classfier to classify seizure vs. non-seizure."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_seizure = (y_train == 1)\n",
    "y_test_seizure = (y_test == 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False    7365\n",
       "True     1835\n",
       "Name: y, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_seizure.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train a `Stochastic Gradient Descent (SGD) Classifier`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n",
       "              early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
       "              l1_ratio=0.15, learning_rate='optimal', loss='hinge',\n",
       "              max_iter=1000, n_iter_no_change=5, n_jobs=None, penalty='l2',\n",
       "              power_t=0.5, random_state=42, shuffle=True, tol=0.001,\n",
       "              validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import SGDClassifier\n",
    "\n",
    "sgd_clf = SGDClassifier(random_state=42)\n",
    "sgd_clf.fit(X_train_prepared, y_train_seizure)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ True])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_data = X_train.iloc[55]\n",
    "\n",
    "sgd_clf.predict([some_data])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_seizure[55]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_predict\n",
    "\n",
    "y_train_predict = cross_val_predict(sgd_clf, X_train, y_train_seizure, cv=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Confusion matrix**:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5576, 1789],\n",
       "       [ 951,  884]], dtype=int64)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix, precision_score, recall_score, f1_score\n",
    "\n",
    "confusion_matrix(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.33071455293677515"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision_score(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.4817438692098093"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.39219165927240457"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_score(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*All 3 scores are under 50%!*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Recall vs. Precision**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "y_scores = cross_val_predict(sgd_clf, X_train, y_train_seizure, cv=3, method=\"decision_function\")\n",
    "precisions, recalls, thresholds = precision_recall_curve(y_train_seizure, y_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(recalls, precisions, label=None)\n",
    "plt.axis([0, 1, 0, 1])\n",
    "plt.xlabel(\"Recall\")\n",
    "plt.ylabel(\"Precision\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(thresholds, precisions[:-1], \"b--\", label=\"Precision\")\n",
    "plt.plot(thresholds, recalls[:-1], \"g-\", label=\"Recall/Sensitivity\")\n",
    "plt.xlabel(\"Threshold\")\n",
    "plt.legend(loc=\"center left\")\n",
    "plt.ylim([0, 1])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**ROC Curve**:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_train_seizure, y_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(fpr, tpr, label=None)\n",
    "plt.plot([0, 1], [0, 1], \"k--\")\n",
    "plt.axis([0, 1, 0, 1])\n",
    "plt.xlabel(\"FPR / 1 - TNR / 1 - specificity\")\n",
    "plt.ylabel(\"TPR / sensitivity\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This model is not performant, try instead a `Random Forest Classifier`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,\n",
       "                       criterion='gini', max_depth=None, max_features='auto',\n",
       "                       max_leaf_nodes=None, max_samples=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "                       n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
       "                       warm_start=False)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "forest_clf = RandomForestClassifier(random_state=42)\n",
    "forest_clf.fit(X_train_prepared, y_train_seizure)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_predict\n",
    "\n",
    "y_train_predict = cross_val_predict(forest_clf, X_train, y_train_seizure, cv=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[7248,  117],\n",
       "       [ 142, 1693]], dtype=int64)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix, precision_score, recall_score, f1_score\n",
    "\n",
    "confusion_matrix(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9353591160220994"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision_score(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9226158038147139"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9289437585733882"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_score(y_train_seizure, y_train_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The performance of this model is much better!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "y_scores = cross_val_predict(forest_clf, X_train, y_train_seizure, cv=3)\n",
    "precisions, recalls, thresholds = precision_recall_curve(y_train_seizure, y_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(recalls, precisions, label=None)\n",
    "plt.axis([0, 1, 0, 1])\n",
    "plt.xlabel(\"Recall\")\n",
    "plt.ylabel(\"Precision\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(thresholds, precisions[:-1], \"b--\", label=\"Precision\")\n",
    "plt.plot(thresholds, recalls[:-1], \"g-\", label=\"Recall/Sensitivity\")\n",
    "plt.xlabel(\"Threshold\")\n",
    "plt.legend(loc=\"center left\")\n",
    "plt.xlim([0, 1])\n",
    "plt.ylim([0, 1])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, auc\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_train_seizure, y_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(fpr, tpr, label=None)\n",
    "plt.plot([0, 1], [0, 1], \"k--\")\n",
    "plt.axis([0, 1, 0, 1])\n",
    "plt.xlabel(\"FPR / 1 - TNR / 1 - specificity\")\n",
    "plt.ylabel(\"TPR / sensitivity\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9533649283839354"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Multiclass Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use the original prepared X training."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use a **Random Forest Classifier** that can directly classify into multiple classes, without using One versus One or One versus Rest strategy:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "forest_clf = RandomForestClassifier(random_state=42)\n",
    "epileptic_seizure_predicted = cross_val_predict(forest_clf, X_train_prepared, \n",
    "                                    y_train, cv=3,\n",
    "                                    method=\"predict_proba\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9200, 5)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "epileptic_seizure_predicted.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fit the training set on the classifier:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,\n",
       "                       criterion='gini', max_depth=None, max_features='auto',\n",
       "                       max_leaf_nodes=None, max_samples=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "                       n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
       "                       warm_start=False)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest_clf.fit(X_train_prepared, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.08, 0.79, 0.04, 0.05, 0.04]])"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_data = X_train_prepared[55]\n",
    "\n",
    "forest_clf.predict_proba([some_data])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Actual `y` value:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[55]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Cross validation**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_predicted = cross_val_predict(forest_clf, X_train, y_train, cv=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**ROC Curves for all classes**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_classes = 5\n",
    "fpr = dict()\n",
    "tpr = dict()\n",
    "roc_auc = dict()\n",
    "\n",
    "for i in range(n_classes):\n",
    "    y_train_i = (y_train == i+1)\n",
    "    y_train_predicted_i = (y_train_predicted == i+1)\n",
    "    fpr[i], tpr[i], _ = roc_curve(y_train_i, y_train_predicted_i)\n",
    "    roc_auc[i] = auc(fpr[i], tpr[i])\n",
    "\n",
    "colors = ['red', 'yellow', 'blue', 'green', 'purple']\n",
    "\n",
    "plt.figure()\n",
    "for i, color in zip(range(n_classes), colors):\n",
    "    plt.plot(fpr[i], tpr[i], color=color, label=\"ROC for class {0} (area= {1:0.2f})\"\n",
    "            ''.format(i+1, roc_auc[i]))\n",
    "    \n",
    "plt.plot([0, 1], [0, 1], \"k--\")\n",
    "plt.axis([0, 1, 0, 1])\n",
    "plt.xlabel(\"FPR / 1 - TNR / 1 - specificity\")\n",
    "plt.ylabel(\"TPR / sensitivity\")\n",
    "plt.legend(loc=\"lower right\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Confusion matrix:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1744,   18,   19,   54,    0],\n",
       "       [ 104,  918,  536,   80,  203],\n",
       "       [  44,  500,  977,  110,  219],\n",
       "       [   9,   36,   42, 1392,  364],\n",
       "       [   0,  144,  230,  323, 1134]], dtype=int64)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conf_matrix = confusion_matrix(y_train, y_train_predicted)\n",
    "conf_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(conf_matrix, cmap=plt.cm.gray)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Class 0 = seizure is best classfied, but the classe 2 is 25% misclassified as class 3 and the class 3 is 20% misclassified as class 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "row_sums = conf_matrix.sum(axis=1, keepdims=True)\n",
    "norm_conf_mx = conf_matrix / row_sums"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "np.fill_diagonal(norm_conf_mx, 0)\n",
    "plt.matshow(norm_conf_mx, cmap=plt.cm.gray)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Predict pipeline**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "predict_pipeline = Pipeline([\n",
    "    ('preparation', num_pipeline),\n",
    "    ('forest_clf', forest_clf)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('preparation',\n",
       "                 Pipeline(memory=None,\n",
       "                          steps=[('std_scaler',\n",
       "                                  StandardScaler(copy=True, with_mean=True,\n",
       "                                                 with_std=True))],\n",
       "                          verbose=False)),\n",
       "                ('forest_clf',\n",
       "                 RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                        class_weight=None, criterion='gini',\n",
       "                                        max_depth=None, max_features='auto',\n",
       "                                        max_leaf_nodes=None, max_samples=None,\n",
       "                                        min_impurity_decrease=0.0,\n",
       "                                        min_impurity_split=None,\n",
       "                                        min_samples_leaf=1, min_samples_split=2,\n",
       "                                        min_weight_fraction_leaf=0.0,\n",
       "                                        n_estimators=100, n_jobs=None,\n",
       "                                        oob_score=False, random_state=42,\n",
       "                                        verbose=0, warm_start=False))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict_pipeline.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Some predictions with test data:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_testing = test_set.drop(\"y\", axis=1) # drop labels for training set\n",
    "y_labels = test_set[\"y\"].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions:\t [3, 4, 4, 5, 5, 2, 3, 2, 5, 1, 3, 5, 5, 1, 5]\n",
      "Labels:\t\t [5, 4, 4, 5, 5, 2, 2, 3, 5, 1, 3, 5, 2, 1, 4]\n"
     ]
    }
   ],
   "source": [
    "some_data = X_test.iloc[:15]\n",
    "some_labels = y_test.iloc[:15]\n",
    "\n",
    "print(\"Predictions:\\t\", list(predict_pipeline.predict(some_data)))\n",
    "print(\"Labels:\\t\\t\", list(some_labels))"
   ]
  }
 ],
 "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": "3.7.6"
  },
  "nav_menu": {
   "height": "279px",
   "width": "309px"
  },
  "toc": {
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}