epileptic_seizure_recognition_ML.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Epileptic seizure recognition"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's start by importing the necessary libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "from sklearn.feature_selection import SelectKBest, RFECV\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis as LDA\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "#from xgboost.sklearn import XGBClassifier\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score, classification_report\n",
    "import seaborn as sn\n",
    "#import keras\n",
    "#from keras.models import Sequential\n",
    "#from keras.layers import Dense, BatchNormalization, Dropout\n",
    "#from keras import regularizers\n",
    "\n",
    "import warnings; warnings.simplefilter('ignore')\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset\n",
    "\n",
    "- The original dataset from the reference consists of 5 different folders, each with 100 files, with each file representing a single subject/person. Each file is a recording of brain activity for 23.6 seconds. \n",
    "\n",
    "- The corresponding time-series is sampled into 4097 data points. Each data point is the value of the EEG recording at a different point in time. So we have total 500 individuals with each has 4097 data points for 23.5 seconds.\n",
    "\n",
    "- We divided and shuffled every 4097 data points into 23 chunks, each chunk contains 178 data points for 1 second, and each data point is the value of the EEG recording at a different point in time. \n",
    "\n",
    "- So now we have 23 x 500 = 11500 pieces of information(row), each information contains 178 data points for 1 second(column), the last column represents the label y {1,2,3,4,5}. \n",
    "\n",
    "- The response variable is y in column 179, the Explanatory variables X1, X2, ..., X178 "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The column y contains the category of the 178-dimensional input vector. \n",
    "\n",
    "### Specifically y in {1, 2, 3, 4, 5}: \n",
    "\n",
    "1. Recording of seizure activity\n",
    "\n",
    "2. They recorded the EEG from the area where the tumor was located\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",
    "4. eyes closed, means when they were recording the EEG signal the patient had their eyes closed \n",
    "\n",
    "5. eyes open, means when they were recording the EEG signal of the brain the patient had their eyes open"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- All subjects falling in classes 2, 3, 4, and 5 are subjects who did not have epileptic seizure. Only subjects in class 1 have epileptic seizure.\n",
    "\n",
    "- Our motivation for creating this version of the data was to simplify access to the data via the creation of a .csv version of it. \n",
    "\n",
    "- Although there are 5 classes most authors have done binary classification, namely class 1 (Epileptic seizure) against the rest.\n",
    "\n",
    "## Citation Request:\n",
    "\n",
    "Andrzejak RG, Lehnertz K, Rieke C, Mormann F, David P, Elger CE (2001) Indications of nonlinear deterministic and finite dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state, Phys. Rev. E, 64, 061907"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing the data\n",
    "\n",
    "Now, let's import the dataset and visualize the first 5 rows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = pd.read_csv('epileptic_seizure_recognition.csv')\n",
    "dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The number of trials for the non-seizure class is: 9200\n",
      "The number of trials for the seizure class is: 2300\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x25938de9278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cols = dataset.columns\n",
    "tgt = dataset.y\n",
    "tgt[tgt>1]=0\n",
    "ax = sn.countplot(tgt,label=\"Count\")\n",
    "non_seizure, seizure = tgt.value_counts()\n",
    "print('The number of trials for the non-seizure class is:', non_seizure)\n",
    "print('The number of trials for the seizure class is:', seizure)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, there are 178 EEG features and 5 possible classes. The main goal of the dataset it's to be able to correctly identify epileptic seizures from EEG data, so a binary classification between classes of label 1 and the rest (2,3,4,5). In order to train our model, let's define our independent variables (X) and our dependent variable (y)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(11500, 178)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = dataset.iloc[:,1:179].values\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'Samples')"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZcAAAEWCAYAAACqitpwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nATVfx9G4xhtIAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2593acfb748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplot(511)\n",
    "plt.plot(X[1,:])\n",
    "plt.title('Classes')\n",
    "plt.ylabel('uV')\n",
    "plt.subplot(512)\n",
    "plt.plot(X[7,:])\n",
    "plt.subplot(513)\n",
    "plt.plot(X[12,:])\n",
    "plt.subplot(514)\n",
    "plt.plot(X[0,:])\n",
    "plt.subplot(515)\n",
    "plt.plot(X[2,:])\n",
    "plt.xlabel('Samples')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = dataset.iloc[:,179].values\n",
    "y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To make this a binary problem, let's turn the non-seizure classes 0 while maintaining the seizure as 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[y>1]=0\n",
    "y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's split our data between training and test. We later will employ 10-fold cross validation to ensure that our model correctly generalize to unseen data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train and test splits"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Applying scaling to our independent variables to make easier for our model to converge."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scaling the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sc = StandardScaler()\n",
    "X_train = sc.fit_transform(X_train)\n",
    "X_test = sc.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.19147932, -0.21477354, -0.21249808, -0.25082442, -0.20864317,\n",
       "       -0.19188163, -0.14740626, -0.1173909 , -0.11400036, -0.14269786,\n",
       "       -0.17140752, -0.18274939, -0.15747181, -0.08557341, -0.03611843,\n",
       "       -0.02916717, -0.0306056 , -0.01310075,  0.02252931,  0.03373113,\n",
       "        0.04517098,  0.00946367, -0.03594922, -0.13818669, -0.19130803,\n",
       "       -0.22796225, -0.23434042, -0.22147656, -0.23909917, -0.26226699])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0,0:30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[0:30]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's investigate how accurate the simplest model can achieve by applying a Logistic Regression to our data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "classifier = LogisticRegression()\n",
    "classifier.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = classifier.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "accuracies = cross_val_score(estimator = classifier,\n",
    "                             X = X_train,\n",
    "                             y = y_train,\n",
    "                             cv = 10,\n",
    "                             n_jobs = -1)\n",
    "accuracies.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cm = confusion_matrix(y_test, y_pred)\n",
    "print(classification_report(y_test, y_pred, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With a simple Logistic Regression the classifier was not able to separate the classes in a linear fashion.\n",
    "\n",
    "Let's try to apply Random Forest with boosting, more specificaly Gradient Boosting, or XGBoost."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "clf = XGBClassifier()\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "accuracies = cross_val_score(estimator = clf,\n",
    "                             X = X_train,\n",
    "                             y = y_train,\n",
    "                             cv = 10,\n",
    "                             n_jobs = -1)\n",
    "accuracies.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "accuracies.std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cm = confusion_matrix(y_test, y_pred)\n",
    "print(classification_report(y_test, y_pred, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using boosting techniques to the random forest classifier (XGBoost) we are able to obtain near perfect accuracies on our data! But that's not telling the whole history, let's investigate our features even further."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Artificial Neural Networks\n",
    "\n",
    "Let's try using a simple, shallow neural network on our problem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9200 samples, validate on 2300 samples\n",
      "Epoch 1/100\n",
      "9200/9200 [==============================] - 3s 348us/step - loss: 0.2635 - acc: 0.9178 - val_loss: 0.1635 - val_acc: 0.9504\n",
      "Epoch 2/100\n",
      "9200/9200 [==============================] - 2s 222us/step - loss: 0.1673 - acc: 0.9476 - val_loss: 0.1493 - val_acc: 0.9470\n",
      "Epoch 3/100\n",
      "9200/9200 [==============================] - 2s 195us/step - loss: 0.1496 - acc: 0.9489 - val_loss: 0.1325 - val_acc: 0.9474\n",
      "Epoch 4/100\n",
      "9200/9200 [==============================] - 2s 173us/step - loss: 0.1370 - acc: 0.9537 - val_loss: 0.1172 - val_acc: 0.9474\n",
      "Epoch 5/100\n",
      "9200/9200 [==============================] - 1s 152us/step - loss: 0.1188 - acc: 0.9558 - val_loss: 0.1110 - val_acc: 0.9517\n",
      "Epoch 6/100\n",
      "9200/9200 [==============================] - 2s 164us/step - loss: 0.1166 - acc: 0.9567 - val_loss: 0.1063 - val_acc: 0.9539\n",
      "Epoch 7/100\n",
      "9200/9200 [==============================] - 1s 162us/step - loss: 0.1120 - acc: 0.9592 - val_loss: 0.0931 - val_acc: 0.9587\n",
      "Epoch 8/100\n",
      "9200/9200 [==============================] - 2s 185us/step - loss: 0.1132 - acc: 0.9570 - val_loss: 0.0981 - val_acc: 0.9596\n",
      "Epoch 9/100\n",
      "9200/9200 [==============================] - 2s 165us/step - loss: 0.1025 - acc: 0.9603 - val_loss: 0.0987 - val_acc: 0.9587\n",
      "Epoch 10/100\n",
      "9200/9200 [==============================] - 2s 170us/step - loss: 0.0971 - acc: 0.9614 - val_loss: 0.0913 - val_acc: 0.9604\n",
      "Epoch 11/100\n",
      "9200/9200 [==============================] - 1s 160us/step - loss: 0.1025 - acc: 0.9602 - val_loss: 0.0854 - val_acc: 0.9648\n",
      "Epoch 12/100\n",
      "9200/9200 [==============================] - 2s 171us/step - loss: 0.0925 - acc: 0.9649 - val_loss: 0.0898 - val_acc: 0.9630\n",
      "Epoch 13/100\n",
      "9200/9200 [==============================] - 1s 146us/step - loss: 0.0958 - acc: 0.9645 - val_loss: 0.0842 - val_acc: 0.9652\n",
      "Epoch 14/100\n",
      "9200/9200 [==============================] - 1s 143us/step - loss: 0.0930 - acc: 0.9653 - val_loss: 0.0835 - val_acc: 0.9670\n",
      "Epoch 15/100\n",
      "9200/9200 [==============================] - 1s 142us/step - loss: 0.0966 - acc: 0.9658 - val_loss: 0.0860 - val_acc: 0.9639\n",
      "Epoch 16/100\n",
      "9200/9200 [==============================] - 1s 142us/step - loss: 0.0898 - acc: 0.9666 - val_loss: 0.0870 - val_acc: 0.9643\n",
      "Epoch 17/100\n",
      "9200/9200 [==============================] - 1s 141us/step - loss: 0.0921 - acc: 0.9650 - val_loss: 0.0852 - val_acc: 0.9652\n",
      "Epoch 18/100\n",
      "9200/9200 [==============================] - 1s 140us/step - loss: 0.0867 - acc: 0.9660 - val_loss: 0.0790 - val_acc: 0.9726\n",
      "Epoch 19/100\n",
      "9200/9200 [==============================] - 1s 142us/step - loss: 0.0895 - acc: 0.9680 - val_loss: 0.0828 - val_acc: 0.9661\n",
      "Epoch 20/100\n",
      "9200/9200 [==============================] - 1s 142us/step - loss: 0.0913 - acc: 0.9673 - val_loss: 0.0847 - val_acc: 0.9665\n",
      "Epoch 21/100\n",
      "9200/9200 [==============================] - 1s 142us/step - loss: 0.0832 - acc: 0.9682 - val_loss: 0.0817 - val_acc: 0.9683\n",
      "Epoch 22/100\n",
      "9200/9200 [==============================] - 1s 142us/step - loss: 0.0802 - acc: 0.9715 - val_loss: 0.0800 - val_acc: 0.9691\n",
      "Epoch 23/100\n",
      "9200/9200 [==============================] - 1s 141us/step - loss: 0.0787 - acc: 0.9699 - val_loss: 0.0772 - val_acc: 0.9713\n",
      "Epoch 24/100\n",
      "9200/9200 [==============================] - 1s 141us/step - loss: 0.0808 - acc: 0.9687 - val_loss: 0.0814 - val_acc: 0.9700\n",
      "Epoch 25/100\n",
      "9200/9200 [==============================] - 1s 141us/step - loss: 0.0759 - acc: 0.9711 - val_loss: 0.0813 - val_acc: 0.9670\n",
      "Epoch 26/100\n",
      "9200/9200 [==============================] - 1s 145us/step - loss: 0.0823 - acc: 0.9704 - val_loss: 0.0847 - val_acc: 0.9657\n",
      "Epoch 27/100\n",
      "9200/9200 [==============================] - 1s 143us/step - loss: 0.0816 - acc: 0.9696 - val_loss: 0.0802 - val_acc: 0.9687\n",
      "Epoch 28/100\n",
      "9200/9200 [==============================] - 1s 142us/step - loss: 0.0803 - acc: 0.9701 - val_loss: 0.0856 - val_acc: 0.9630\n",
      "Epoch 00028: early stopping\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "Non-seizure       0.97      0.99      0.98      1827\n",
      "    Seizure       0.94      0.87      0.91       473\n",
      "\n",
      "avg / total       0.96      0.96      0.96      2300\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f85567c17f0>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWcAAAD8CAYAAACrbmW5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npZDD2cwshRzOZmYp5HA2M0uh/wf8flBmBHOBWwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f85567c1518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "early_stop = keras.callbacks.EarlyStopping(monitor='val_acc',\n",
    "                                           min_delta=0,\n",
    "                                           patience=10,\n",
    "                                           verbose=1,\n",
    "                                           mode='auto')\n",
    "# Initialising the ANN\n",
    "clf_ann = Sequential()\n",
    "\n",
    "# Adding the input layer and the first hidden layer\n",
    "clf_ann.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\", \n",
    "                  units=100,\n",
    "                  input_dim=178))\n",
    "clf_ann.add(BatchNormalization())\n",
    "clf_ann.add(Dropout(0.5))\n",
    "# Adding the second hidden layer\n",
    "clf_ann.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  units=100))\n",
    "clf_ann.add(BatchNormalization())\n",
    "clf_ann.add(Dropout(0.5))\n",
    "# Adding the output layer\n",
    "clf_ann.add(Dense(units=1,\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  activation=\"sigmoid\"))\n",
    "\n",
    "# Compiling the ANN\n",
    "clf_ann.compile(optimizer = 'adam',\n",
    "                loss = 'binary_crossentropy',\n",
    "                metrics = ['accuracy'])\n",
    "# loss = 'categorical_crossentropy'\n",
    "# Fitting the ANN to the Training set\n",
    "clf_ann.fit(X_train, y_train,\n",
    "               batch_size = 32,\n",
    "               epochs = 100,\n",
    "               validation_data=(X_test,y_test),\n",
    "               callbacks=[early_stop])\n",
    "\n",
    "# Part 3 - Making the predictions and evaluating the model\n",
    "\n",
    "# Predicting the Test set results\n",
    "y_pred = clf_ann.predict(X_test)\n",
    "y_pred = (y_pred > 0.5)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "cm = confusion_matrix(y_test, y_pred)\n",
    "print(classification_report(y_test, y_pred, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# find best scored 5 features\n",
    "select_feature = SelectKBest(k=10).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "x_train_2 = select_feature.transform(X_train)\n",
    "x_test_2 = select_feature.transform(X_test)\n",
    "#random forest classifier with n_estimators=10 (default)\n",
    "clf_2 = XGBClassifier()     \n",
    "clr_2 = clf_2.fit(x_train_2,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = clf_2.predict(x_test_2)\n",
    "\n",
    "accuracies = cross_val_score(estimator = clf_2,\n",
    "                             X = x_train_2,\n",
    "                             y = y_train,\n",
    "                             cv = 10,\n",
    "                             n_jobs = -1)\n",
    "accuracies.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "accuracies.std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# The \"accuracy\" scoring is proportional to the number of correct classifications\n",
    "clf_3 = RandomForestClassifier() \n",
    "rfecv = RFECV(estimator=clf_3, step=1, cv=5,scoring='accuracy', n_jobs=-1)   #5-fold cross-validation\n",
    "rfecv = rfecv.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "plt.figure()\n",
    "plt.xlabel(\"Number of features selected\")\n",
    "plt.ylabel(\"Cross validation score of number of selected features\")\n",
    "plt.plot(range(1, len(rfecv.grid_scores_) + 1), rfecv.grid_scores_)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Principal Component Analysis (PCA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pca = PCA(n_components = 2)\n",
    "X_train_pca = pca.fit_transform(X_train)\n",
    "X_test_pca = pca.transform(X_test)\n",
    "explained_variance = pca.explained_variance_ratio_\n",
    "explained_variance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.sum(explained_variance)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see PCA is not a good technique for our dataset, we need at least 20 components to have ~70% of explained variance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear Discriminant Analysis (LDA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Applying LDA\n",
    "lda = LDA(n_components = 2)\n",
    "X_train_lda = lda.fit_transform(X_train, y_train)\n",
    "X_test_lda = lda.transform(X_test)\n",
    "\n",
    "# Fitting Logistic Regression to the Training set\n",
    "classifier = LogisticRegression()\n",
    "classifier.fit(X_train_lda, y_train)\n",
    "\n",
    "# Predicting the Test set results\n",
    "y_pred = classifier.predict(X_test_lda)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "cm = confusion_matrix(y_test, y_pred)\n",
    "print(classification_report(y_test, y_pred, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once again, linear predictors don't show good separation between classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature enginnering\n",
    "\n",
    "We are using the time domain features of the EEG in our classification. The literature shows a lot of studies that proves that there are a lot of important information on the frequency domain.\n",
    "\n",
    "Let's apply the Fourier Transform to our dataset and use the frequency domain features on our classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 135,  190,  229,  223,  192,  125,   55,   -9,  -33,  -38,  -10,\n",
       "         35,   64,  113,  152,  164,  127,   50,  -47, -121, -138, -125,\n",
       "       -101,  -50,   11,   39,   24,   48,   64,   46])"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_new = dataset.iloc[:,1:179].values\n",
    "X_new[0,0:30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
       "       1, 1, 1, 0, 0, 1, 0, 0])"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_fft = dataset.iloc[:,179].values\n",
    "y_fft[0:30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(11500, 90)"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_fft = np.fft.rfft(X_new,axis=1)\n",
    "X_fft.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f853f631a20>]"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY8AAAD8CAYAAACPWyg8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8546fa1f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplot(511)\n",
    "plt.plot(X_fft[1,:])\n",
    "plt.subplot(512)\n",
    "plt.plot(X_fft[7,:])\n",
    "plt.subplot(513)\n",
    "plt.plot(X_fft[12,:])\n",
    "plt.subplot(514)\n",
    "plt.plot(X_fft[0,:])\n",
    "plt.subplot(515)\n",
    "plt.plot(X_fft[2,:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As it's evident, it was applied a lowpass filter at 40 Hz (this information is also given by the dataset owners).\n",
    "\n",
    "So let's use only the first 40 points of our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f853f4bc048>]"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY8AAAD8CAYAAACPWyg8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8546fa1208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_fft = X_fft[:,0:30]\n",
    "plt.subplot(511)\n",
    "plt.plot(X_fft[1,:])\n",
    "plt.subplot(512)\n",
    "plt.plot(X_fft[7,:])\n",
    "plt.subplot(513)\n",
    "plt.plot(X_fft[12,:])\n",
    "plt.subplot(514)\n",
    "plt.plot(X_fft[0,:])\n",
    "plt.subplot(515)\n",
    "plt.plot(X_fft[2,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,\n",
       "       max_depth=3, min_child_weight=1, missing=None, n_estimators=100,\n",
       "       n_jobs=1, nthread=None, objective='binary:logistic', random_state=0,\n",
       "       reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
       "       silent=True, subsample=1)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_fft, X_test_fft, y_train_fft, y_test_fft = train_test_split(X_fft, y, test_size = 0.2)\n",
    "sc = StandardScaler()\n",
    "X_train_fft = sc.fit_transform(X_train_fft)\n",
    "X_test_fft = sc.transform(X_test_fft)\n",
    "clf_fft = XGBClassifier()\n",
    "clf_fft.fit(X_train_fft, y_train_fft)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.964565665324767"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_fft = clf_fft.predict(X_test_fft)\n",
    "accuracies = cross_val_score(estimator = clf_fft,\n",
    "                             X = X_train_fft,\n",
    "                             y = y_train_fft,\n",
    "                             cv = 10,\n",
    "                             n_jobs = -1)\n",
    "accuracies.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.005304384138896117"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracies.std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "Non-seizure       0.97      0.99      0.98      1823\n",
      "    Seizure       0.96      0.88      0.92       477\n",
      "\n",
      "avg / total       0.97      0.97      0.97      2300\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8556813748>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWcAAAD8CAYAAACrbmW5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8556826400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test_fft, y_pred_fft)\n",
    "print(classification_report(y_test_fft, y_pred_fft, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Time + frequency features\n",
    "\n",
    "Let's see if using both features at the same time it's possible to achieve a better accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(11500, 208)"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_fft = np.fft.rfft(X_new,axis=1)\n",
    "X_fft = X_fft[:,0:30]\n",
    "X_both = np.concatenate((X_new,X_fft),axis=1)\n",
    "X_both.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,\n",
       "       max_depth=3, min_child_weight=1, missing=None, n_estimators=100,\n",
       "       n_jobs=1, nthread=None, objective='binary:logistic', random_state=0,\n",
       "       reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
       "       silent=True, subsample=1)"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_both, X_test_both, y_train_both, y_test_both = train_test_split(X_both, y, test_size = 0.2)\n",
    "sc = StandardScaler()\n",
    "X_train_both = sc.fit_transform(X_train_both)\n",
    "X_test_both = sc.transform(X_test_both)\n",
    "clf_both = XGBClassifier()\n",
    "clf_both.fit(X_train_both, y_train_both)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9689130368003538"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_both = clf_both.predict(X_test_both)\n",
    "accuracies = cross_val_score(estimator = clf_both,\n",
    "                             X = X_train_both,\n",
    "                             y = y_train_both,\n",
    "                             cv = 10,\n",
    "                             n_jobs = -1)\n",
    "accuracies.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0042433267676053234"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracies.std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "Non-seizure       0.97      0.99      0.98      1831\n",
      "    Seizure       0.96      0.87      0.91       469\n",
      "\n",
      "avg / total       0.97      0.97      0.97      2300\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8543e240b8>"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWcAAAD8CAYAAACrbmW5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8543eeb5c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test_both, y_pred_both)\n",
    "print(classification_report(y_test_both, y_pred_both, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9200 samples, validate on 2300 samples\n",
      "Epoch 1/100\n",
      "9200/9200 [==============================] - 5s 525us/step - loss: 0.2597 - acc: 0.9195 - val_loss: 0.1343 - val_acc: 0.9596\n",
      "Epoch 2/100\n",
      "9200/9200 [==============================] - 2s 222us/step - loss: 0.1769 - acc: 0.9425 - val_loss: 0.1166 - val_acc: 0.9565\n",
      "Epoch 3/100\n",
      "9200/9200 [==============================] - 2s 206us/step - loss: 0.1523 - acc: 0.9482 - val_loss: 0.1075 - val_acc: 0.9561\n",
      "Epoch 4/100\n",
      "9200/9200 [==============================] - 2s 217us/step - loss: 0.1309 - acc: 0.9536 - val_loss: 0.0953 - val_acc: 0.9622\n",
      "Epoch 5/100\n",
      "9200/9200 [==============================] - 2s 221us/step - loss: 0.1272 - acc: 0.9537 - val_loss: 0.0870 - val_acc: 0.9652\n",
      "Epoch 6/100\n",
      "9200/9200 [==============================] - 2s 255us/step - loss: 0.1112 - acc: 0.9578 - val_loss: 0.0864 - val_acc: 0.9661\n",
      "Epoch 7/100\n",
      "9200/9200 [==============================] - 2s 231us/step - loss: 0.1152 - acc: 0.9572 - val_loss: 0.0758 - val_acc: 0.9726\n",
      "Epoch 8/100\n",
      "9200/9200 [==============================] - 2s 176us/step - loss: 0.1129 - acc: 0.9580 - val_loss: 0.0775 - val_acc: 0.9717\n",
      "Epoch 9/100\n",
      "9200/9200 [==============================] - 2s 174us/step - loss: 0.1041 - acc: 0.9628 - val_loss: 0.0722 - val_acc: 0.9735\n",
      "Epoch 10/100\n",
      "9200/9200 [==============================] - 2s 181us/step - loss: 0.1023 - acc: 0.9630 - val_loss: 0.0742 - val_acc: 0.9704\n",
      "Epoch 11/100\n",
      "9200/9200 [==============================] - 2s 192us/step - loss: 0.0978 - acc: 0.9633 - val_loss: 0.0696 - val_acc: 0.9743\n",
      "Epoch 12/100\n",
      "9200/9200 [==============================] - 2s 189us/step - loss: 0.0974 - acc: 0.9648 - val_loss: 0.0697 - val_acc: 0.9726\n",
      "Epoch 13/100\n",
      "9200/9200 [==============================] - 2s 188us/step - loss: 0.0959 - acc: 0.9649 - val_loss: 0.0744 - val_acc: 0.9704\n",
      "Epoch 14/100\n",
      "9200/9200 [==============================] - 2s 198us/step - loss: 0.0919 - acc: 0.9655 - val_loss: 0.0690 - val_acc: 0.9726\n",
      "Epoch 15/100\n",
      "9200/9200 [==============================] - 2s 181us/step - loss: 0.0918 - acc: 0.9667 - val_loss: 0.0690 - val_acc: 0.9748\n",
      "Epoch 16/100\n",
      "9200/9200 [==============================] - 2s 194us/step - loss: 0.0885 - acc: 0.9687 - val_loss: 0.0674 - val_acc: 0.9739\n",
      "Epoch 17/100\n",
      "9200/9200 [==============================] - 2s 203us/step - loss: 0.0886 - acc: 0.9653 - val_loss: 0.0677 - val_acc: 0.9743\n",
      "Epoch 18/100\n",
      "9200/9200 [==============================] - 2s 215us/step - loss: 0.0922 - acc: 0.9664 - val_loss: 0.0701 - val_acc: 0.9730\n",
      "Epoch 19/100\n",
      "9200/9200 [==============================] - 2s 190us/step - loss: 0.0881 - acc: 0.9682 - val_loss: 0.0668 - val_acc: 0.9752\n",
      "Epoch 20/100\n",
      "9200/9200 [==============================] - 2s 190us/step - loss: 0.0898 - acc: 0.9659 - val_loss: 0.0707 - val_acc: 0.9752\n",
      "Epoch 21/100\n",
      "9200/9200 [==============================] - 2s 187us/step - loss: 0.0847 - acc: 0.9688 - val_loss: 0.0648 - val_acc: 0.9761\n",
      "Epoch 22/100\n",
      "9200/9200 [==============================] - 2s 193us/step - loss: 0.0797 - acc: 0.9708 - val_loss: 0.0637 - val_acc: 0.9783\n",
      "Epoch 23/100\n",
      "9200/9200 [==============================] - 2s 167us/step - loss: 0.0813 - acc: 0.9698 - val_loss: 0.0646 - val_acc: 0.9774\n",
      "Epoch 24/100\n",
      "9200/9200 [==============================] - 2s 165us/step - loss: 0.0780 - acc: 0.9714 - val_loss: 0.0621 - val_acc: 0.9796\n",
      "Epoch 25/100\n",
      "9200/9200 [==============================] - 2s 167us/step - loss: 0.0830 - acc: 0.9673 - val_loss: 0.0628 - val_acc: 0.9796\n",
      "Epoch 26/100\n",
      "9200/9200 [==============================] - 2s 200us/step - loss: 0.0768 - acc: 0.9709 - val_loss: 0.0605 - val_acc: 0.9778\n",
      "Epoch 27/100\n",
      "9200/9200 [==============================] - 2s 229us/step - loss: 0.0768 - acc: 0.9730 - val_loss: 0.0615 - val_acc: 0.9778\n",
      "Epoch 28/100\n",
      "9200/9200 [==============================] - 2s 201us/step - loss: 0.0795 - acc: 0.9692 - val_loss: 0.0634 - val_acc: 0.9783\n",
      "Epoch 29/100\n",
      "9200/9200 [==============================] - 2s 249us/step - loss: 0.0754 - acc: 0.9722 - val_loss: 0.0622 - val_acc: 0.9813\n",
      "Epoch 30/100\n",
      "9200/9200 [==============================] - 2s 175us/step - loss: 0.0810 - acc: 0.9693 - val_loss: 0.0637 - val_acc: 0.9765\n",
      "Epoch 31/100\n",
      "9200/9200 [==============================] - 2s 206us/step - loss: 0.0765 - acc: 0.9716 - val_loss: 0.0607 - val_acc: 0.9804\n",
      "Epoch 32/100\n",
      "9200/9200 [==============================] - 2s 177us/step - loss: 0.0684 - acc: 0.9742 - val_loss: 0.0655 - val_acc: 0.9770\n",
      "Epoch 33/100\n",
      "9200/9200 [==============================] - 2s 166us/step - loss: 0.0777 - acc: 0.9713 - val_loss: 0.0653 - val_acc: 0.9761\n",
      "Epoch 34/100\n",
      "9200/9200 [==============================] - 2s 169us/step - loss: 0.0701 - acc: 0.9745 - val_loss: 0.0644 - val_acc: 0.9757\n",
      "Epoch 35/100\n",
      "9200/9200 [==============================] - 2s 168us/step - loss: 0.0665 - acc: 0.9762 - val_loss: 0.0596 - val_acc: 0.9804\n",
      "Epoch 36/100\n",
      "9200/9200 [==============================] - 2s 176us/step - loss: 0.0667 - acc: 0.9747 - val_loss: 0.0602 - val_acc: 0.9787\n",
      "Epoch 37/100\n",
      "9200/9200 [==============================] - 2s 164us/step - loss: 0.0730 - acc: 0.9736 - val_loss: 0.0587 - val_acc: 0.9770\n",
      "Epoch 38/100\n",
      "9200/9200 [==============================] - 2s 165us/step - loss: 0.0716 - acc: 0.9730 - val_loss: 0.0619 - val_acc: 0.9778\n",
      "Epoch 39/100\n",
      "9200/9200 [==============================] - 2s 182us/step - loss: 0.0686 - acc: 0.9742 - val_loss: 0.0630 - val_acc: 0.9787\n",
      "Epoch 00039: early stopping\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "Non-seizure       0.98      0.99      0.99      1844\n",
      "    Seizure       0.96      0.93      0.95       456\n",
      "\n",
      "avg / total       0.98      0.98      0.98      2300\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8542a4ddd8>"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWcAAAD8CAYAAACrbmW5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSaSA2lIhhnIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8542a4d978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "early_stop = keras.callbacks.EarlyStopping(monitor='val_acc',\n",
    "                                           min_delta=0,\n",
    "                                           patience=10,\n",
    "                                           verbose=1,\n",
    "                                           mode='auto')\n",
    "# Initialising the ANN\n",
    "clf_ann2 = Sequential()\n",
    "\n",
    "# Adding the input layer and the first hidden layer\n",
    "clf_ann2.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\", \n",
    "                  units=100,\n",
    "                  input_dim=208))\n",
    "clf_ann2.add(BatchNormalization())\n",
    "clf_ann2.add(Dropout(0.5))\n",
    "# Adding the second hidden layer\n",
    "clf_ann2.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  units=100))\n",
    "clf_ann2.add(BatchNormalization())\n",
    "clf_ann2.add(Dropout(0.5))\n",
    "\n",
    "# Adding the output layer\n",
    "clf_ann2.add(Dense(units=1,\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  activation=\"sigmoid\"))\n",
    "\n",
    "# Compiling the ANN\n",
    "clf_ann2.compile(optimizer = 'adam',\n",
    "                loss = 'binary_crossentropy',\n",
    "                metrics = ['accuracy'])\n",
    "# loss = 'categorical_crossentropy'\n",
    "# Fitting the ANN to the Training set\n",
    "clf_ann2.fit(X_train_both, \n",
    "            y_train_both,\n",
    "            batch_size = 32,\n",
    "            epochs = 100,\n",
    "            validation_data=(X_test_both,y_test_both),\n",
    "            callbacks=[early_stop])\n",
    "\n",
    "# Part 3 - Making the predictions and evaluating the model\n",
    "\n",
    "# Predicting the Test set results\n",
    "y_pred_both = clf_ann2.predict(X_test_both)\n",
    "y_pred_both = (y_pred_both > 0.5)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "cm = confusion_matrix(y_test_both, y_pred_both)\n",
    "print(classification_report(y_test_both, y_pred_both, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Oversampling our seizure class\n",
    "\n",
    "As we saw in the beginning of this notebook, our classes are imbalanced.\n",
    "\n",
    "- 2300 seizure trials (20%)\n",
    "- 9200 non-seizure trials (80%)\n",
    "\n",
    "This could lead to a common problem in classification: the accuracy paradox.\n",
    "If our classifier chooses to go all in on the non-seizure class, it still achieves 80% accuracy!\n",
    "\n",
    "We dealt with this problem nicely by investigating other metrics of our classifier and our model seems to generalize well."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SMOTE\n",
    "\n",
    "Other technique used to deal with imbalanced classes is oversampling. Let's investigate on our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE\n",
    "X_s = dataset.iloc[:,1:179].values\n",
    "y_s = dataset.iloc[:,179].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14664, 178)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_s, X_test_s, y_train_s, y_test_s = train_test_split(X_s, y_s,test_size = 0.2)\n",
    "sm = SMOTE(ratio = 1.0)\n",
    "X_train_res, y_train_res = sm.fit_sample(X_train_s, y_train_s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 14664 samples, validate on 2300 samples\n",
      "Epoch 1/100\n",
      "14664/14664 [==============================] - 4s 273us/step - loss: 0.2322 - acc: 0.9120 - val_loss: 0.1085 - val_acc: 0.9652\n",
      "Epoch 2/100\n",
      "14664/14664 [==============================] - 3s 218us/step - loss: 0.1915 - acc: 0.9277 - val_loss: 0.1058 - val_acc: 0.9626\n",
      "Epoch 3/100\n",
      "14664/14664 [==============================] - 3s 222us/step - loss: 0.1824 - acc: 0.9321 - val_loss: 0.0956 - val_acc: 0.9709\n",
      "Epoch 4/100\n",
      "14664/14664 [==============================] - 3s 227us/step - loss: 0.1799 - acc: 0.9331 - val_loss: 0.1159 - val_acc: 0.9587\n",
      "Epoch 5/100\n",
      "14664/14664 [==============================] - 3s 222us/step - loss: 0.1861 - acc: 0.9294 - val_loss: 0.0991 - val_acc: 0.9674\n",
      "Epoch 6/100\n",
      "14664/14664 [==============================] - 3s 224us/step - loss: 0.1775 - acc: 0.9334 - val_loss: 0.1076 - val_acc: 0.9643\n",
      "Epoch 7/100\n",
      "14664/14664 [==============================] - 3s 222us/step - loss: 0.1654 - acc: 0.9394 - val_loss: 0.0843 - val_acc: 0.9691\n",
      "Epoch 8/100\n",
      "14664/14664 [==============================] - 3s 223us/step - loss: 0.1649 - acc: 0.9367 - val_loss: 0.0918 - val_acc: 0.9635\n",
      "Epoch 9/100\n",
      "14664/14664 [==============================] - 3s 224us/step - loss: 0.1640 - acc: 0.9381 - val_loss: 0.0912 - val_acc: 0.9691\n",
      "Epoch 10/100\n",
      "14664/14664 [==============================] - 3s 230us/step - loss: 0.1652 - acc: 0.9365 - val_loss: 0.0876 - val_acc: 0.9674\n",
      "Epoch 11/100\n",
      "14664/14664 [==============================] - 3s 228us/step - loss: 0.1583 - acc: 0.9384 - val_loss: 0.0855 - val_acc: 0.9670\n",
      "Epoch 12/100\n",
      "14664/14664 [==============================] - 3s 223us/step - loss: 0.1633 - acc: 0.9356 - val_loss: 0.0909 - val_acc: 0.9661\n",
      "Epoch 13/100\n",
      "14664/14664 [==============================] - 3s 231us/step - loss: 0.1634 - acc: 0.9393 - val_loss: 0.0925 - val_acc: 0.9704\n",
      "Epoch 00013: early stopping\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "Non-seizure       0.99      0.97      0.98      1868\n",
      "    Seizure       0.89      0.97      0.92       432\n",
      "\n",
      "avg / total       0.97      0.97      0.97      2300\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f61c8fb56d8>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWcAAAD8CAYAAACrbmW5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f61c8fb5d30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "early_stop = keras.callbacks.EarlyStopping(monitor='val_acc',\n",
    "                                           min_delta=0,\n",
    "                                           patience=10,\n",
    "                                           verbose=1,\n",
    "                                           mode='auto')\n",
    "# Initialising the ANN\n",
    "clf_ann = Sequential()\n",
    "\n",
    "# Adding the input layer and the first hidden layer\n",
    "clf_ann.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\", \n",
    "                  units=100,\n",
    "                  input_dim=178))\n",
    "clf_ann.add(BatchNormalization())\n",
    "clf_ann.add(Dropout(0.5))\n",
    "# Adding the second hidden layer\n",
    "clf_ann.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  units=100))\n",
    "clf_ann.add(BatchNormalization())\n",
    "clf_ann.add(Dropout(0.5))\n",
    "# Adding the second hidden layer\n",
    "clf_ann.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  units=100))\n",
    "clf_ann.add(BatchNormalization())\n",
    "clf_ann.add(Dropout(0.5))\n",
    "# Adding the output layer\n",
    "clf_ann.add(Dense(units=1,\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  activation=\"sigmoid\"))\n",
    "\n",
    "# Compiling the ANN\n",
    "clf_ann.compile(optimizer = 'adam',\n",
    "                loss = 'binary_crossentropy',\n",
    "                metrics = ['accuracy'])\n",
    "# loss = 'categorical_crossentropy'\n",
    "# Fitting the ANN to the Training set\n",
    "clf_ann.fit(X_train_res, y_train_res,\n",
    "               batch_size = 32,\n",
    "               epochs = 100,\n",
    "               validation_data=(X_test_s,y_test_s),\n",
    "               callbacks=[early_stop])\n",
    "\n",
    "# Part 3 - Making the predictions and evaluating the model\n",
    "\n",
    "# Predicting the Test set results\n",
    "y_pred_s = clf_ann.predict(X_test_s)\n",
    "y_pred_s = (y_pred_s > 0.5)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "cm = confusion_matrix(y_test_s, y_pred_s)\n",
    "print(classification_report(y_test_s, y_pred_s, target_names=['Non-seizure', 'Seizure']))\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(2), range(2))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predicting 5 classes\n",
    "\n",
    "Now, let's transform the problem in a 5 class classification, this time we wont transform our targets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = pd.read_csv('epileptic_seizure_recognition.csv')\n",
    "dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 135,  190,  229,  223,  192,  125,   55,   -9,  -33,  -38,  -10,\n",
       "         35,   64,  113,  152,  164,  127,   50,  -47, -121, -138, -125,\n",
       "       -101,  -50,   11,   39,   24,   48,   64,   46])"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = dataset.iloc[:,1:179].values\n",
    "X[0,0:30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4, 1, 5, 5, 5, 5, 4, 2, 1, 4, 5, 1, 3, 4, 2, 3, 2, 3, 4, 2, 1, 2,\n",
       "       1, 1, 1, 3, 5, 1, 2, 2])"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = dataset.iloc[:,179].values\n",
    "y[0:30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2300 2300 2300 2300 2300\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAD5dJREFUeJzt3X3MnXV9x/H3hwJzQw0lvcGuLatZ\nGrPuCVhTyZqo04WnqWVmGEmUzrHUP8BpZrbg/hhOQmKy6aboSJhW6eYkZMisWyM2HZPohtBq5VHT\nxjHo2tFKfZyJBvzuj3PVHu3du+dXes51bu73K7lz7vPrdU6/nD9491zXda6TqkKSpFGd0vcAkqT5\nxXBIkpoYDklSE8MhSWpiOCRJTQyHJKmJ4ZAkNTEckqQmhkOS1OTUvgcYhyVLltTKlSv7HkOS5pWd\nO3d+o6pmjrfdczIcK1euZMeOHX2PIUnzSpL/HmU7d1VJkpoYDklSE8MhSWpiOCRJTQyHJKmJ4ZAk\nNTEckqQmhkOS1MRwSJKaPCc/OT6X3/iTzX2PMBY7//Kq5sc8/u5fHcMk/Tv3zx9sfsy6m9aNYZL+\nfeGtX2h+zOde9vIxTNK/l9/zuebHfPAdnx7DJP279r2veVaP9x2HJKmJ4ZAkNTEckqQmhkOS1MRw\nSJKaGA5JUhPDIUlqYjgkSU0MhySpieGQJDUxHJKkJoZDktTEcEiSmhgOSVITwyFJamI4JElNDIck\nqYnhkCQ1MRySpCaGQ5LUxHBIkpoYDklSk7GFI8mKJHcneTTJw0ne1q2flWRbkt3d7eJuPUk+kGRP\nkgeSXDD0XBu67Xcn2TCumSVJxzfOdxxPA++oql8CLgSuSbIauA7YXlWrgO3dfYBLgVXdz0bgZhiE\nBrgeeCmwFrj+cGwkSZM3tnBU1f6q+lL3+3eBR4FlwHrg1m6zW4HLu9/XA5tr4F7gzCRLgYuBbVV1\nqKq+CWwDLhnX3JKkuU3kGEeSlcD5wBeBc6pqPwziApzdbbYMeGLoYXu7tWOtS5J6MPZwJHk+cAfw\n9qr6zlybzrJWc6z/9N+zMcmOJDsOHjx4YsNKko5rrOFIchqDaHy8qj7ZLT/Z7YKiuz3Qre8FVgw9\nfDmwb471n1BVt1TVmqpaMzMzc3L/QyRJPzbOs6oCfAR4tKreN/RHW4DDZ0ZtAD41tH5Vd3bVhcC3\nu11ZdwEXJVncHRS/qFuTJPXg1DE+9zrgTcCDSXZ1a38GvAe4PcnVwOPAFd2fbQUuA/YA3wfeDFBV\nh5LcANzfbffuqjo0xrklSXMYWziq6vPMfnwC4FWzbF/ANcd4rk3AppM3nSTpRPnJcUlSE8MhSWpi\nOCRJTQyHJKmJ4ZAkNTEckqQmhkOS1MRwSJKaGA5JUhPDIUlqYjgkSU0MhySpieGQJDUxHJKkJoZD\nktTEcEiSmhgOSVITwyFJamI4JElNDIckqYnhkCQ1MRySpCaGQ5LUxHBIkpoYDklSE8MhSWpiOCRJ\nTQyHJKmJ4ZAkNTEckqQmhkOS1MRwSJKaGA5JUhPDIUlqYjgkSU0MhySpydjCkWRTkgNJHhpae1eS\n/0myq/u5bOjP3plkT5KvJbl4aP2Sbm1PkuvGNa8kaTTjfMfxMeCSWdb/uqrO6362AiRZDbwB+OXu\nMX+bZFGSRcCHgEuB1cCV3baSpJ6cOq4nrqp7kqwccfP1wG1V9QPgv5LsAdZ2f7anqr4OkOS2bttH\nTvK4kqQR9XGM49okD3S7shZ3a8uAJ4a22dutHWtdktSTSYfjZuAXgfOA/cB7u/XMsm3NsX6UJBuT\n7Eiy4+DBgydjVknSLCYajqp6sqqeqaofAX/Hkd1Re4EVQ5suB/bNsT7bc99SVWuqas3MzMzJH16S\nBEw4HEmWDt39XeDwGVdbgDck+ZkkLwZWAfcB9wOrkrw4yekMDqBvmeTMkqSfNLaD40k+AbwCWJJk\nL3A98Iok5zHY3fQY8BaAqno4ye0MDno/DVxTVc90z3MtcBewCNhUVQ+Pa2ZJ0vGN86yqK2dZ/sgc\n298I3DjL+lZg60kcTZL0LPjJcUlSE8MhSWpiOCRJTQyHJKmJ4ZAkNTEckqQmhkOS1MRwSJKaGA5J\nUpORwpFk+yhrkqTnvjkvOZLkecDPMbje1GKOXOb8hcDPj3k2SdIUOt61qt4CvJ1BJHZyJBzfYfCV\nrpKkBWbOcFTV+4H3J3lrVd00oZkkSVNspKvjVtVNSX4TWDn8mKraPKa5JElTaqRwJPl7Bl/5ugt4\nplsuwHBI0gIz6vdxrAFWV9Ws3/ctSVo4Rv0cx0PAi8Y5iCRpfhj1HccS4JEk9wE/OLxYVa8dy1SS\npKk1ajjeNc4hJEnzx6hnVX1u3INIkuaHUc+q+i6Ds6gATgdOA/6vql44rsEkSdNp1HccLxi+n+Ry\nYO1YJpIkTbUTujpuVf0z8MqTPIskaR4YdVfV64bunsLgcx1+pkOSFqBRz6p6zdDvTwOPAetP+jSS\npKk36jGON497EEnS/DDqFzktT3JnkgNJnkxyR5Ll4x5OkjR9Rj04/lFgC4Pv5VgGfLpbkyQtMKOG\nY6aqPlpVT3c/HwNmxjiXJGlKjRqObyR5Y5JF3c8bgafGOZgkaTqNGo4/AF4P/C+wH/g9wAPmkrQA\njXo67g3Ahqr6JkCSs4C/YhAUSdICMuo7jl87HA2AqjoEnD+ekSRJ02zUcJySZPHhO907jlHfrUiS\nnkNG/Z//e4H/SPJPDC418nrgxrFNJUmaWqN+cnxzkh0MLmwY4HVV9chYJ5MkTaWRr45bVY9U1Qer\n6qZRopFkU/dJ84eG1s5Ksi3J7u52cbeeJB9IsifJA0kuGHrMhm773Uk2tP4HSpJOrhO6rPqIPgZc\n8lNr1wHbq2oVsL27D3ApsKr72QjcDD8+lnI98FIG3/9x/fCxFknS5I0tHFV1D3Dop5bXA7d2v98K\nXD60vrkG7gXOTLIUuBjYVlWHurO6tnF0jCRJEzTOdxyzOaeq9gN0t2d368uAJ4a229utHWtdktST\nSYfjWDLLWs2xfvQTJBuT7Eiy4+DBgyd1OEnSEZMOx5PdLii62wPd+l5gxdB2y4F9c6wfpapuqao1\nVbVmZsbrL0rSuEw6HFuAw2dGbQA+NbR+VXd21YXAt7tdWXcBFyVZ3B0Uv6hbkyT1ZGyf/k7yCeAV\nwJIkexmcHfUe4PYkVwOPA1d0m28FLgP2AN+nu4BiVR1KcgNwf7fdu7vLnUiSejK2cFTVlcf4o1fN\nsm0B1xzjeTYBm07iaJKkZ2FaDo5LkuYJwyFJamI4JElNDIckqYnhkCQ1MRySpCaGQ5LUxHBIkpoY\nDklSE8MhSWpiOCRJTQyHJKmJ4ZAkNTEckqQmhkOS1MRwSJKaGA5JUhPDIUlqYjgkSU0MhySpieGQ\nJDUxHJKkJoZDktTEcEiSmhgOSVITwyFJamI4JElNDIckqYnhkCQ1MRySpCaGQ5LUxHBIkpoYDklS\nE8MhSWpiOCRJTQyHJKlJL+FI8liSB5PsSrKjWzsrybYku7vbxd16knwgyZ4kDyS5oI+ZJUkDfb7j\n+K2qOq+q1nT3rwO2V9UqYHt3H+BSYFX3sxG4eeKTSpJ+bJp2Va0Hbu1+vxW4fGh9cw3cC5yZZGkf\nA0qS+gtHAZ9NsjPJxm7tnKraD9Ddnt2tLwOeGHrs3m5NktSDU3v6e9dV1b4kZwPbknx1jm0zy1od\ntdEgQBsBzj333JMzpSTpKL2846iqfd3tAeBOYC3w5OFdUN3tgW7zvcCKoYcvB/bN8py3VNWaqloz\nMzMzzvElaUGbeDiSnJHkBYd/By4CHgK2ABu6zTYAn+p+3wJc1Z1ddSHw7cO7tCRJk9fHrqpzgDuT\nHP77/7GqPpPkfuD2JFcDjwNXdNtvBS4D9gDfB948+ZElSYdNPBxV9XXg12dZfwp41SzrBVwzgdEk\nSSOYptNxJUnzgOGQJDUxHJKkJoZDktTEcEiSmhgOSVITwyFJamI4JElNDIckqYnhkCQ1MRySpCaG\nQ5LUxHBIkpoYDklSE8MhSWpiOCRJTQyHJKmJ4ZAkNTEckqQmhkOS1MRwSJKaGA5JUhPDIUlqYjgk\nSU0MhySpieGQJDUxHJKkJoZDktTEcEiSmhgOSVITwyFJamI4JElNDIckqYnhkCQ1MRySpCaGQ5LU\nZN6EI8klSb6WZE+S6/qeR5IWqnkRjiSLgA8BlwKrgSuTrO53KklamOZFOIC1wJ6q+npV/RC4DVjf\n80yStCDNl3AsA54Yur+3W5MkTViqqu8ZjivJFcDFVfWH3f03AWur6q1D22wENnZ3XwJ8beKDHm0J\n8I2+h5gSvhZH+Foc4WtxxDS8Fr9QVTPH2+jUSUxyEuwFVgzdXw7sG96gqm4BbpnkUMeTZEdVrel7\njmnga3GEr8URvhZHzKfXYr7sqrofWJXkxUlOB94AbOl5JklakObFO46qejrJtcBdwCJgU1U93PNY\nkrQgzYtwAFTVVmBr33M0mqpdZz3ztTjC1+IIX4sj5s1rMS8OjkuSpsd8OcYhSZoShmMMkmxKciDJ\nQ33P0qckK5LcneTRJA8neVvfM/UlyfOS3JfkK91r8Rd9z9S3JIuSfDnJv/Q9S5+SPJbkwSS7kuzo\ne55RuKtqDJK8DPgesLmqfqXvefqSZCmwtKq+lOQFwE7g8qp6pOfRJi5JgDOq6ntJTgM+D7ytqu7t\nebTeJPljYA3wwqp6dd/z9CXJY8Caqur7Mxwj8x3HGFTVPcChvufoW1Xtr6ovdb9/F3iUBfqJ/xr4\nXnf3tO5nwf6rLcly4HeAD/c9i9oZDk1EkpXA+cAX+52kP92umV3AAWBbVS3Y1wL4G+BPgR/1PcgU\nKOCzSXZ2V8CYeoZDY5fk+cAdwNur6jt9z9OXqnqmqs5jcOWDtUkW5G7MJK8GDlTVzr5nmRLrquoC\nBlf/vqbb1T3VDIfGqtuffwfw8ar6ZN/zTIOq+hbw78AlPY/Sl3XAa7t9+7cBr0zyD/2O1J+q2tfd\nHgDuZHA18KlmODQ23QHhjwCPVtX7+p6nT0lmkpzZ/f6zwG8DX+13qn5U1TuranlVrWRw+aB/q6o3\n9jxWL5Kc0Z04QpIzgIuAqT8b03CMQZJPAP8JvCTJ3iRX9z1TT9YBb2LwL8pd3c9lfQ/Vk6XA3Uke\nYHDttW1VtaBPQxUA5wCfT/IV4D7gX6vqMz3PdFyejitJauI7DklSE8MhSWpiOCRJTQyHJKmJ4ZAk\nNTEckqQmhkOS1MRwSBOQ5Ibh7yNJcmOSP+pzJulE+QFAaQK6qwN/sqouSHIKsBtYW1VP9TqYdAJO\n7XsAaSGoqseSPJXkfAaXmfiy0dB8ZTikyfkw8PvAi4BN/Y4inTh3VUkTkuR04EEG3/63qqqe6Xkk\n6YT4jkOakKr6YZK7gW8ZDc1nhkOakO6g+IXAFX3PIj0bno4rTUCS1cAeYHtV7e57HunZ8BiHJKmJ\n7zgkSU0MhySpieGQJDUxHJKkJoZDktTEcEiSmvw/+trVi6GB5CgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8542b229b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tgt = dataset.y\n",
    "ax = sn.countplot(tgt,label=\"Count\")\n",
    "a,b,c,d,e = tgt.value_counts()\n",
    "print(a,b,c,d,e)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sc = StandardScaler()\n",
    "X_train = sc.fit_transform(X_train)\n",
    "X_test = sc.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "classifier = XGBClassifier()\n",
    "classifier.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = classifier.predict(X_test)\n",
    "\n",
    "accuracies = cross_val_score(estimator = classifier,\n",
    "                             X = X_train,\n",
    "                             y = y_train,\n",
    "                             cv = 10,\n",
    "                             n_jobs = -1)\n",
    "accuracies.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "accuracies.std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cm = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "print(classification_report(y_test, y_pred, target_names=['Seizure',\n",
    "                                                          'Tumor Area',\n",
    "                                                          'Healthy Area',\n",
    "                                                          'Eyes closed',\n",
    "                                                          'Eyes opened']))\n",
    "\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(5), range(5))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = dataset.iloc[:,1:179].values\n",
    "y = dataset.iloc[:,179].values\n",
    "y = y-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<11500x5 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 11500 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import scipy.sparse\n",
    "indptr = range(len(y)+1)\n",
    "data = np.ones(len(y))\n",
    "y_ann = scipy.sparse.csr_matrix((data, y, indptr))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y_ann, test_size = 0.2)\n",
    "sc = StandardScaler()\n",
    "X_train = sc.fit_transform(X_train)\n",
    "X_test = sc.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9200 samples, validate on 2300 samples\n",
      "Epoch 1/100\n",
      "9200/9200 [==============================] - 6s 651us/step - loss: 1.2506 - acc: 0.4320 - val_loss: 1.0878 - val_acc: 0.5591\n",
      "Epoch 2/100\n",
      "9200/9200 [==============================] - 2s 256us/step - loss: 1.1119 - acc: 0.5003 - val_loss: 1.0065 - val_acc: 0.5861\n",
      "Epoch 3/100\n",
      "9200/9200 [==============================] - 2s 267us/step - loss: 1.0497 - acc: 0.5270 - val_loss: 0.9494 - val_acc: 0.6087\n",
      "Epoch 4/100\n",
      "9200/9200 [==============================] - 2s 269us/step - loss: 1.0104 - acc: 0.5400 - val_loss: 0.9431 - val_acc: 0.5835\n",
      "Epoch 5/100\n",
      "9200/9200 [==============================] - 2s 272us/step - loss: 0.9941 - acc: 0.5496 - val_loss: 0.9064 - val_acc: 0.6217\n",
      "Epoch 6/100\n",
      "9200/9200 [==============================] - 3s 276us/step - loss: 0.9733 - acc: 0.5566 - val_loss: 0.9376 - val_acc: 0.5761\n",
      "Epoch 7/100\n",
      "9200/9200 [==============================] - 2s 272us/step - loss: 0.9631 - acc: 0.5649 - val_loss: 0.9010 - val_acc: 0.6013\n",
      "Epoch 8/100\n",
      "9200/9200 [==============================] - 3s 272us/step - loss: 0.9485 - acc: 0.5705 - val_loss: 0.9477 - val_acc: 0.5700\n",
      "Epoch 9/100\n",
      "9200/9200 [==============================] - 3s 289us/step - loss: 0.9355 - acc: 0.5771 - val_loss: 0.9254 - val_acc: 0.5904\n",
      "Epoch 10/100\n",
      "9200/9200 [==============================] - 3s 284us/step - loss: 0.9357 - acc: 0.5773 - val_loss: 0.9590 - val_acc: 0.5648\n",
      "Epoch 11/100\n",
      "9200/9200 [==============================] - 3s 293us/step - loss: 0.9333 - acc: 0.5724 - val_loss: 0.9294 - val_acc: 0.5683\n",
      "Epoch 12/100\n",
      "9200/9200 [==============================] - 2s 268us/step - loss: 0.9300 - acc: 0.5702 - val_loss: 0.9379 - val_acc: 0.5804\n",
      "Epoch 13/100\n",
      "9200/9200 [==============================] - 3s 286us/step - loss: 0.9079 - acc: 0.5808 - val_loss: 0.9125 - val_acc: 0.5865\n",
      "Epoch 14/100\n",
      "9200/9200 [==============================] - 3s 273us/step - loss: 0.9106 - acc: 0.5837 - val_loss: 0.9675 - val_acc: 0.5761\n",
      "Epoch 15/100\n",
      "9200/9200 [==============================] - 3s 284us/step - loss: 0.8987 - acc: 0.5821 - val_loss: 0.9126 - val_acc: 0.5917\n",
      "Epoch 00015: early stopping\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "     Seizure       0.96      0.84      0.89       482\n",
      "  Tumor Area       1.00      0.01      0.02       460\n",
      "Healthy Area       0.00      0.00      0.00       458\n",
      " Eyes closed       0.88      0.63      0.73       446\n",
      " Eyes opened       0.75      0.02      0.04       454\n",
      "\n",
      " avg / total       0.72      0.30      0.34      2300\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f853a8950b8>"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAD8CAYAAABJsn7AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkf8PvfdrMQoAJkMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f853a8957f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "early_stop = keras.callbacks.EarlyStopping(monitor='val_acc',\n",
    "                                           min_delta=0,\n",
    "                                           patience=10,\n",
    "                                           verbose=1,\n",
    "                                           mode='auto')\n",
    "# Initialising the ANN\n",
    "clf_ann3 = Sequential()\n",
    "\n",
    "# Adding the input layer and the first hidden layer\n",
    "clf_ann3.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\", \n",
    "                  units=100,\n",
    "                  input_dim=178))\n",
    "clf_ann3.add(BatchNormalization())\n",
    "clf_ann3.add(Dropout(0.5))\n",
    "# Adding the second hidden layer\n",
    "clf_ann3.add(Dense(activation=\"relu\",\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  units=100))\n",
    "clf_ann3.add(BatchNormalization())\n",
    "clf_ann3.add(Dropout(0.5))\n",
    "\n",
    "# Adding the output layer\n",
    "clf_ann3.add(Dense(units=5,\n",
    "                  kernel_initializer=\"uniform\",\n",
    "                  activation=\"softmax\"))\n",
    "\n",
    "# Compiling the ANN\n",
    "clf_ann3.compile(optimizer = 'adam',\n",
    "                loss = 'categorical_crossentropy',\n",
    "                metrics = ['accuracy'])\n",
    "# loss = 'categorical_crossentropy'\n",
    "# Fitting the ANN to the Training set\n",
    "clf_ann3.fit(X_train, \n",
    "            y_train,\n",
    "            batch_size = 32,\n",
    "            epochs = 100,\n",
    "            validation_data=(X_test,y_test),\n",
    "            callbacks=[early_stop])\n",
    "\n",
    "# Part 3 - Making the predictions and evaluating the model\n",
    "\n",
    "# Predicting the Test set results\n",
    "y_pred = clf_ann3.predict(X_test)\n",
    "y_pred = (y_pred > 0.5)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "cm = confusion_matrix(y_test.argmax(axis=1), y_pred.argmax(axis=1))\n",
    "print(classification_report(y_test, y_pred, target_names=['Seizure',\n",
    "                                                          'Tumor Area',\n",
    "                                                          'Healthy Area',\n",
    "                                                          'Eyes closed',\n",
    "                                                          'Eyes opened']))\n",
    "\n",
    "\n",
    "df_cm = pd.DataFrame(cm, range(5), range(5))\n",
    "sn.heatmap(df_cm, annot=True,fmt='g',cmap ='Blues')# font size"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}