weighted_Subject av_dictionary based classification without BP and CSP_final.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "jWidL5EtmRyU"
   },
   "outputs": [],
   "source": [
    "#Five fold cross validation\n",
    "\n",
    "\n",
    "import numpy as np\n",
    "import pyedflib\n",
    "from matplotlib import pyplot as plt\n",
    "from nitime import utils\n",
    "from nitime import algorithms as alg\n",
    "from nitime.timeseries import TimeSeries\n",
    "from nitime.viz import plot_tseries\n",
    "import csv\n",
    "import pywt\n",
    "import scipy.stats as sp\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis as LDA\n",
    "from scipy import signal\n",
    "from spectrum import *\n",
    "from os import listdir\n",
    "from os.path import isfile, join\n",
    "from wyrm import processing as proc\n",
    "from wyrm.types import Data\n",
    "from wyrm.io import convert_mushu_data\n",
    "from sklearn import metrics\n",
    "from wyrm.processing import calculate_csp,segment_dat,apply_csp,append_epo\n",
    "from wyrm.processing import select_channels\n",
    "from wyrm.processing import swapaxes\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import OrthogonalMatchingPursuit\n",
    "from sklearn.linear_model import OrthogonalMatchingPursuitCV\n",
    "from sklearn.preprocessing import normalize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "GiKlkcmlmRyh"
   },
   "outputs": [],
   "source": [
    "files1  = [f for f in listdir('/home/singh/projects/BCI/binary/Training data')]\n",
    "\n",
    "\n",
    "channels = ['Fp1', 'AFp1', 'Fpz', 'AFp2', 'Fp2', 'AF7', 'AF3', 'AF4', 'AF8', 'FAF5', 'FAF1', 'FAF2', 'FAF6', \n",
    "                'F7', 'F5', 'F3', 'F1', 'Fz', 'F2', 'F4', 'F6', 'F8', 'FFC7', 'FFC5', 'FFC3', 'FFC1', 'FFC2', 'FFC4', \n",
    "                'FFC6', 'FFC8', 'FT9', 'FT7', 'FC5', 'FC3', 'FC1', 'FCz', 'FC2', 'FC4', 'FC6', 'FT8', 'FT10', 'CFC7', \n",
    "                'CFC5', 'CFC3', 'CFC1', 'CFC2', 'CFC4', 'CFC6', 'CFC8', 'T7', 'C5', 'C3', 'C1', 'Cz', 'C2', 'C4', 'C6'\n",
    "                , 'T8', 'CCP7', 'CCP5', 'CCP3', 'CCP1', 'CCP2', 'CCP4', 'CCP6', 'CCP8', 'TP9', 'TP7', 'CP5', 'CP3', 'CP1',\n",
    "                'CPz', 'CP2', 'CP4', 'CP6', 'TP8', 'TP10', 'PCP7', 'PCP5', 'PCP3', 'PCP1', 'PCP2', 'PCP4', 'PCP6', 'PCP8', \n",
    "                'P9', 'P7', 'P5', 'P3', 'P1', 'Pz', 'P2', 'P4', 'P6', 'P8', 'P10', 'PPO7', 'PPO5', 'PPO1', 'PPO2', 'PPO6',\n",
    "                'PPO8', 'PO7', 'PO3', 'PO1', 'POz', 'PO2', 'PO4', 'PO8', 'OPO1', 'OPO2', 'O1', 'Oz', 'O2', 'OI1', 'OI2', \n",
    "                'I1', 'I2']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "RunFse04mRyp",
    "outputId": "cd385a42-60f8-472b-809f-d282f9cf346c"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ay', 'aw', 'aa', 'al', 'av']"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "files1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "rqwQfIBmmRy0"
   },
   "outputs": [],
   "source": [
    "#for 1st subject\n",
    "sub = 3\n",
    "training_data = '/home/singh/projects/BCI/binary/Training data/' + files1[sub] + '/data_set_IVa_' + files1[sub] + '_cnt.txt'\n",
    "markers       = '/home/singh/projects/BCI/binary/Training data/' + files1[sub] + '/data_set_IVa_' + files1[sub] + '_mrk.txt'\n",
    "    \n",
    "signal_array = [map(float,l.split()) for l in open(training_data).readlines()]\n",
    "    \n",
    "signal_array = np.array(signal_array)\n",
    "    \n",
    "marker_array = [map(str,l.split()) for l in open(markers).readlines()]\n",
    "        \n",
    "time = np.arange(len(signal_array))\n",
    "    \n",
    "train_markers1 = [(float(events[0]),str(events[1])) for events in marker_array if events[1]!= 0]\n",
    "\n",
    "\n",
    "for events in marker_array:\n",
    "    if events[1] != 0:\n",
    "        train_markers1.append((float(events[0]) + 50.0, str(events[1])))\n",
    "        train_markers1.append((float(events[0]) + 100.0, str(events[1])))\n",
    "        train_markers1.append((float(events[0]) + 150.0, str(events[1])))\n",
    "        train_markers1.append((float(events[0]) + 200.0, str(events[1])))\n",
    "        train_markers1.append((float(events[0]) + 250.0, str(events[1])))\n",
    "        train_markers1.append((float(events[0]) + 300.0, str(events[1])))\n",
    "        train_markers1.append((float(events[0]) + 350.0, str(events[1])))\n",
    "\n",
    "markers1 = np.array(train_markers1)\n",
    "\n",
    "y1 = [0] * len(markers1)\n",
    "\n",
    "\n",
    "markers_subject1_class_1 = [(float(events[0]),str(events[1])) for events in markers1 if events[1]== '1']\n",
    "markers_subject1_class_2 = [(float(events[0]),str(events[1])) for events in markers1 if events[1]== '2']\n",
    "\n",
    "\n",
    "cnt1 = convert_mushu_data(signal_array, markers_subject1_class_1,50,channels)\n",
    "cnt2 = convert_mushu_data(signal_array, markers_subject1_class_2,50,channels)\n",
    "    \n",
    "md = {'class 1': ['1'],'class 2': ['2']}\n",
    "    \n",
    "epoch_subject1_class1 = segment_dat(cnt1, md, [0, 1000])\n",
    "epoch_subject1_class2 = segment_dat(cnt2, md, [0, 1000])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "7ydnD8VqmRy4"
   },
   "outputs": [],
   "source": [
    "# Appending all the different subject training epochs\n",
    "\n",
    "final_epoch = append_epo(epoch_subject1_class1,epoch_subject1_class2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "FxZj0kDymRy6",
    "outputId": "acf38127-f9c5-4660-9bac-256e4a44a6a7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1792, 50, 118)\n"
     ]
    }
   ],
   "source": [
    "print (final_epoch.data.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "6zog3oRqmRy_"
   },
   "outputs": [],
   "source": [
    "####### FUNCTION TO FIND BAND POWERS ###########\n",
    "\n",
    "def bandpowers(segment):\n",
    "    features = []\n",
    "    for i in range(len(segment)):\n",
    "        f,Psd = signal.welch(segment[i,:], 100)\n",
    "        power1 = 0\n",
    "        power2 = 0\n",
    "        f1 = []\n",
    "        for j in range(0,len(f)):\n",
    "            if(f[j]>=4 and f[j]<=13):\n",
    "                power1 += Psd[j]\n",
    "            if(f[j]>=14 and f[j]<=30):\n",
    "                power2 += Psd[j]\n",
    "        features.append(power1)\n",
    "        features.append(power2)\n",
    "    return features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "QOtA4RHdmRzD"
   },
   "outputs": [],
   "source": [
    "######### Discrete Cosine Transform\n",
    "from scipy.fftpack import fft, dct\n",
    "def dct_features(segment):\n",
    "    features = []\n",
    "    for i in range(len(segment)):\n",
    "        dct_coef = dct(segment[i,:], 2, norm='ortho')\n",
    "        power = sum( j*j for j in dct_coef)\n",
    "        features.append(power)\n",
    "    return features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "DV_FCEsZmRzF"
   },
   "outputs": [],
   "source": [
    "####### FUNCTION TO FIND WAVELET FEATURES#######\n",
    "\n",
    "def wavelet_features(epoch):\n",
    "    cA_values = []\n",
    "    cD_values = []\n",
    "    cA_mean = []\n",
    "    cA_std = []\n",
    "    cA_Energy =[]\n",
    "    cD_mean = []\n",
    "    cD_std = []\n",
    "    cD_Energy = []\n",
    "    Entropy_D = []\n",
    "    Entropy_A = []\n",
    "    features = []\n",
    "    for i in range(len(epoch)):\n",
    "        cA,cD=pywt.dwt(epoch[i,:],'coif1')\n",
    "        cA_values.append(cA)\n",
    "        cD_values.append(cD)\t\t#calculating the coefficients of wavelet transform.\n",
    "    for x in range(len(epoch)):   \n",
    "        cA_Energy.append(abs(np.sum(np.square(cA_values[x]))))\n",
    "        features.append(abs(np.sum(np.square(cA_values[x]))))\n",
    "        \n",
    "    for x in range(len(epoch)):      \n",
    "        cD_Energy.append(abs(np.sum(np.square(cD_values[x]))))\n",
    "        features.append(abs(np.sum(np.square(cD_values[x]))))\n",
    "        \n",
    "    return features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "W63Ook7rmRzI",
    "outputId": "5f772453-6212-408f-b415-81271c5e3eaa"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/singh/miniconda3/envs/data2/lib/python2.7/site-packages/scipy/signal/spectral.py:1773: UserWarning: nperseg = 256 is greater than input length  = 50, using nperseg = 50\n",
      "  .format(nperseg, input_length))\n"
     ]
    }
   ],
   "source": [
    "## FULL FEATURES DICTIONARY\n",
    "\n",
    "# Dictionary with band power features\n",
    "#form the dictionary using bandpower features where each column is a data point.\n",
    "\n",
    "#For class 1\n",
    "dictionary1 = []\n",
    "dictionary2 = []\n",
    "dictionary3 = []\n",
    "for i in range(len(final_epoch.axes[0])):\n",
    "    segment = final_epoch.data[i]\n",
    "    segment = np.array(segment)\n",
    "    segment = np.transpose(segment)\n",
    "    \n",
    "   \n",
    "    features1 = bandpowers(segment)  ## Comment the dictionaries not used\n",
    "    #features = logvariance(segment)\n",
    "    \n",
    "    features2 = wavelet_features(segment)\n",
    "    features3 = dct_features(segment)\n",
    "    \n",
    "    dictionary1.append(features1)\n",
    "    dictionary2.append(features2)\n",
    "    dictionary3.append(features3)\n",
    "    \n",
    "    \n",
    "dictionary1 = np.array(dictionary1)\n",
    "dictionary2 = np.array(dictionary2)\n",
    "dictionary3 = np.array(dictionary3)\n",
    "\n",
    "\n",
    "\n",
    "dictionary_bandpower  = dictionary1\n",
    "dictionary_wavelet = dictionary2\n",
    "dictionary_dct = dictionary3\n",
    "\n",
    "\n",
    "targets = final_epoch.axes[0]\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "CDbx22f7mRzM",
    "outputId": "52c5bfe1-3646-48ba-9844-bab94466e8e2"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1792, 236)"
      ]
     },
     "execution_count": 15,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dictionary_wavelet.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "d_XaBTQhmRzP"
   },
   "outputs": [],
   "source": [
    "#FOr wavelet\n",
    "\n",
    "from scipy.sparse import coo_matrix\n",
    "X_sparse = coo_matrix(dictionary_wavelet)\n",
    "\n",
    "from sklearn.utils import resample\n",
    "dictionary_wavelet, X_sparse, ywe = resample(dictionary_wavelet, X_sparse,targets , random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "luaCPq2RmRzR"
   },
   "outputs": [],
   "source": [
    "#For Band Power\n",
    "\n",
    "from scipy.sparse import coo_matrix\n",
    "X_sparse = coo_matrix(dictionary_bandpower)\n",
    "\n",
    "from sklearn.utils import resample\n",
    "dictionary_bandpower, X_sparse, ybp = resample(dictionary_bandpower, X_sparse,targets , random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "F6KanuNImRzT"
   },
   "outputs": [],
   "source": [
    "#For DCT\n",
    "\n",
    "from scipy.sparse import coo_matrix\n",
    "X_sparse = coo_matrix(dictionary_dct)\n",
    "\n",
    "from sklearn.utils import resample\n",
    "dictionary_dct, X_sparse, y = resample(dictionary_dct, X_sparse,targets , random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "z0RIOFGOmRzU",
    "outputId": "336646a5-d429-458b-b2c7-2d570a9aa8a8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold(n_splits=10, random_state=30, shuffle=True)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import KFold\n",
    "\n",
    "#dictionary = dictionary_bandpower\n",
    "dictionary = dictionary_wavelet\n",
    "#dictionary = dictionary_dct\n",
    "\n",
    "kf = KFold(n_splits=10,random_state = 30, shuffle = True)\n",
    "kf.get_n_splits(dictionary)\n",
    "\n",
    "y_classifier1 = []\n",
    "y_classifier2 = []\n",
    "y_classifier3 = []\n",
    "y_classifier4 = []\n",
    "\n",
    "\n",
    "y_all1 = []\n",
    "y_all2 = []\n",
    "y_all3 = []\n",
    "y_all4 = []\n",
    "y_example_test = []\n",
    "\n",
    "print(kf)\n",
    "\n",
    "\n",
    "\n",
    "y_final_test = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "U2X1QMjdmRzX"
   },
   "outputs": [],
   "source": [
    "\n",
    "def calculate_accuracy(Xts,yts,D,class1,class2,n):\n",
    "    y_pred1 = []\n",
    "    y_pred2 = []\n",
    "    y_pred3 = []\n",
    "    y_pred4 = []\n",
    "    diff = []\n",
    "    counter1 = 0\n",
    "    counter2 = 0\n",
    "    \n",
    "    sigma_1 = metrics.pairwise.euclidean_distances(dictionary_wavelet.transpose())\n",
    "    sigma = np.sum(sigma_1)/((dictionary_wavelet.shape[1])*(dictionary_wavelet.shape[1]))\n",
    "    \n",
    "    for i in range(len(Xts)):\n",
    "        features = Xts[i]\n",
    "        \n",
    "        # Find Eucledian distance between single test case with all atoms \n",
    "        # in dictionary and store in vector(Weights) of size equal to number of test cases.\n",
    "        Weights = np.zeros((D.shape[1],))\n",
    "        #for j in range(D.shape[1]):\n",
    "         #   Weights[j] = spatial.distance.euclidean(features, D.transpose()[j])\n",
    "        \n",
    "        Weights = metrics.pairwise.euclidean_distances(D.transpose(), features.reshape(1,-1)).flatten()\n",
    "        \n",
    "        # Apply formula w_i = exp(-1*dist(y,d_i)/sigma) on vector, elementwise.\n",
    "        # Subtract from maximum to normalize.\n",
    "        weights = np.exp(-1*Weights/sigma)\n",
    "        weights = np.amax(weights) - weights\n",
    "        \n",
    "        # Get new weighted dictionary by multiplying weights to D.\n",
    "        D_dash = (D)*weights\n",
    "        \n",
    "        # Apply OMP with new dictionary (using old code)\n",
    "        omp = OrthogonalMatchingPursuit(n_nonzero_coefs=3)\n",
    "        omp.fit(D_dash,features)\n",
    "        coef = omp.coef_\n",
    "        \n",
    "        p = 0\n",
    "        q = 0\n",
    "        l = 0\n",
    "        m = 0\n",
    "        a = 0\n",
    "        b = 0\n",
    "        \n",
    "                \n",
    "        list1 = coef[0:min(class1,class2)]\n",
    "        list2 = coef[min(class1,class2)+1:2*min(class1,class2)]\n",
    "        \n",
    "        c1 = (sum(z*z for z in list1))**(1/2.0) \n",
    "        c2 = (sum(z*z for z in list2))**(1/2.0) \n",
    "        \n",
    "        p = np.std(list1)\n",
    "        q = np.std(list2)\n",
    "        \n",
    "        a = max(list1)\n",
    "        b = max(list2)\n",
    "        \n",
    "        for ko in range(min(class1,class2)):\n",
    "            l = l + coef[ko]\n",
    "        \n",
    "        for io in xrange(min(class1,class2)+1,2*min(class1,class2)):\n",
    "            m = m + coef[io]\n",
    "            \n",
    "        if p > q:\n",
    "            y_pred1.append(0)\n",
    "          \n",
    "        else:\n",
    "            y_pred1.append(1)\n",
    "            if(yts[i] != 0):\n",
    "                if(counter1==0):\n",
    "                    counter1 +=1\n",
    "                    idx_r, = coef.nonzero()\n",
    "                    #plt.xlim(0, len(coef))\n",
    "                    #plt.title(\"Sparse Signal\")\n",
    "                    #plt.stem(idx_r, coef[idx_r])\n",
    "                    #plt.show()\n",
    "            \n",
    "        if l>m:\n",
    "            y_pred2.append(0)\n",
    "        else:\n",
    "            y_pred2.append(1)\n",
    "        \n",
    "        if a>b:\n",
    "            y_pred3.append(0)\n",
    "        else:\n",
    "            y_pred3.append(1)\n",
    "            \n",
    "        if c1 > c2:\n",
    "            y_pred4.append(0)\n",
    "        else:\n",
    "            y_pred4.append(1)\n",
    "            \n",
    "    p1 = metrics.accuracy_score(yts, y_pred4, normalize=True, sample_weight=None)\n",
    "    p2 = metrics.accuracy_score(yts, y_pred3, normalize=True, sample_weight=None)\n",
    "    p3 = metrics.accuracy_score(yts, y_pred1, normalize=True, sample_weight=None)\n",
    "    p4 = metrics.accuracy_score(yts, y_pred2, normalize=True, sample_weight=None)\n",
    "    print('class1', p1)  \n",
    "    \n",
    "    print('class2', p2) \n",
    "    \n",
    "    print('class3', p3)\n",
    "    \n",
    "    print('class4', p4)\n",
    "    \n",
    "    print('\\n')\n",
    "    \n",
    "    \n",
    "    return y_pred4,y_pred3,y_pred1,y_pred2, p1, p2, p3, p4\n",
    "    \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "3pLRsCpzmRzY",
    "outputId": "9dd77abf-3d8c-492b-8915-cdccd4274c4a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold(n_splits=10, random_state=50, shuffle=True)\n",
      "('class1', 0.98529411764705888)\n",
      "('class2', 0.98529411764705888)\n",
      "('class3', 0.98529411764705888)\n",
      "('class4', 0.98529411764705888)\n",
      "\n",
      "\n",
      "('class1', 0.98529411764705888)\n",
      "('class2', 0.98529411764705888)\n",
      "('class3', 0.98529411764705888)\n",
      "('class4', 0.98529411764705888)\n",
      "\n",
      "\n",
      "('class1', 1.0)\n",
      "('class2', 1.0)\n",
      "('class3', 1.0)\n",
      "('class4', 1.0)\n",
      "\n",
      "\n",
      "('class1', 1.0)\n",
      "('class2', 1.0)\n",
      "('class3', 1.0)\n",
      "('class4', 1.0)\n",
      "\n",
      "\n",
      "('class1', 0.9850746268656716)\n",
      "('class2', 0.9850746268656716)\n",
      "('class3', 0.9850746268656716)\n",
      "('class4', 0.9850746268656716)\n",
      "\n",
      "\n",
      "('class1', 0.97014925373134331)\n",
      "('class2', 0.97014925373134331)\n",
      "('class3', 0.97014925373134331)\n",
      "('class4', 0.97014925373134331)\n",
      "\n",
      "\n",
      "('class1', 0.9850746268656716)\n",
      "('class2', 0.9850746268656716)\n",
      "('class3', 0.9850746268656716)\n",
      "('class4', 0.9850746268656716)\n",
      "\n",
      "\n",
      "('class1', 0.94029850746268662)\n",
      "('class2', 0.94029850746268662)\n",
      "('class3', 0.94029850746268662)\n",
      "('class4', 0.94029850746268662)\n",
      "\n",
      "\n",
      "('class1', 1.0)\n",
      "('class2', 1.0)\n",
      "('class3', 1.0)\n",
      "('class4', 1.0)\n",
      "\n",
      "\n",
      "('class1', 0.95522388059701491)\n",
      "('class2', 0.95522388059701491)\n",
      "('class3', 0.95522388059701491)\n",
      "('class4', 0.95522388059701491)\n",
      "\n",
      "\n",
      "('class1_avg', 0.98064091308165047)\n",
      "('class2_avg', 0.98064091308165047)\n",
      "('class3_avg', 0.98064091308165047)\n",
      "('class4_avg', 0.98064091308165047)\n",
      "--- 1.95636892319 seconds ---\n"
     ]
    }
   ],
   "source": [
    "## K fold cross validation\n",
    "import time\n",
    "from sklearn.model_selection import KFold\n",
    "\n",
    "#dictionary = dictionary_bandpower\n",
    "dictionary = dictionary_wavelet\n",
    "#dictionary = dictionary_dct\n",
    "\n",
    "kf = KFold(n_splits=10, random_state=50, shuffle=True)\n",
    "kf.get_n_splits(dictionary)\n",
    "\n",
    "y_classifier1 = []\n",
    "y_classifier2 = []\n",
    "y_classifier3 = []\n",
    "y_classifier4 = []\n",
    "\n",
    "y_all1 = []\n",
    "y_all2 = []\n",
    "y_all3 = []\n",
    "y_all4 = []\n",
    "y_example_test = []\n",
    "\n",
    "psum1 = 0\n",
    "psum2 = 0\n",
    "psum3 = 0\n",
    "psum4 = 0\n",
    "\n",
    "y_final_test = []\n",
    "\n",
    "print(kf)\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "for train_index, test_index in kf.split(dictionary):\n",
    "    X_train, X_test = dictionary[train_index], dictionary[test_index]\n",
    "    y_train, y_test = ywe[train_index], ywe[test_index]\n",
    "    class1 = 0\n",
    "    class2 = 0\n",
    "    for i in range(len(y_train)):\n",
    "        if (y_train[i] == 0):\n",
    "            class1 += 1\n",
    "        else:\n",
    "            class2 += 1\n",
    "    reb_y = []\n",
    "    reb_dic = []\n",
    "    count = 0\n",
    "    iterator = 0\n",
    "    while (count < min(class1, class2)):\n",
    "        if y_train[iterator] == 0:\n",
    "            reb_dic.append(X_train[iterator])\n",
    "            reb_y.append(0)\n",
    "            count += 1\n",
    "        iterator += 1\n",
    "\n",
    "    count = 0\n",
    "    iterator = 0\n",
    "    while (count < min(class1, class2)):\n",
    "        if y_train[iterator] == 1:\n",
    "            reb_dic.append(X_train[iterator])\n",
    "            reb_y.append(1)\n",
    "            count += 1\n",
    "        iterator += 1\n",
    "    reb_dictionary = np.array(reb_dic)\n",
    "    reb_dictionary = reb_dictionary.transpose()\n",
    "    y_classifier1, y_classifier2, y_classifier3, y_classifier4, p1, p2, p3, p4 = calculate_accuracy(\n",
    "        X_test, y_test, reb_dictionary, class1, class2, 5)\n",
    "    psum1 = psum1 + p1\n",
    "    psum2 = psum2 + p2\n",
    "    psum3 = psum3 + p3\n",
    "    psum4 = psum4 + p4\n",
    "\n",
    "    y_all1.extend(y_classifier1)\n",
    "    y_all2.extend(y_classifier2)\n",
    "    y_all3.extend(y_classifier3)\n",
    "    y_all4.extend(y_classifier4)\n",
    "\n",
    "    y_final_test.extend(y_test)\n",
    "    y_example_test = y_test\n",
    "print('class1_avg', psum1/10.0)\n",
    "print('class2_avg', psum2/10.0)\n",
    "print('class3_avg', psum3/10.0)\n",
    "print('class4_avg', psum4/10.0)\n",
    "print(\"--- %s seconds ---\" % (time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "_AduXq6XmRza",
    "outputId": "273b8b05-83fa-4cd9-bdf3-2253bce20bfe"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('class1', 1.0)\n",
      "('class2', 1.0)\n",
      "('class3', 1.0)\n",
      "('class4', 1.0)\n",
      "\n",
      "\n",
      "--- 0.00912499427795 seconds ---\n"
     ]
    }
   ],
   "source": [
    "# Test case running time for 1 test case\n",
    "start_time = time.time()\n",
    "y_classifier1,y_classifier2,y_classifier3,y_classifier4 ,a1,a2,a3,a4= calculate_accuracy(np.reshape(X_test[0],(1,-1)), np.reshape(y_test[0],(1,-1)),reb_dictionary,class1,class2,5)\n",
    "print(\"--- %s seconds ---\" % (time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "lc_l6VN8mRzd"
   },
   "outputs": [],
   "source": [
    "import itertools\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn import svm, datasets\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Greens):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "    print(cm)\n",
    "\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, cm[i, j],\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "vDEcv5pmmRze",
    "outputId": "d2eb0574-75c2-4976-bcfb-f00285d9840d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[645   2]\n",
      " [  4 693]]\n",
      "Normalized confusion matrix\n",
      "[[ 1.    0.  ]\n",
      " [ 0.01  0.99]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVcAAAEmCAYAAADWT9N8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x14eefe750>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVcAAAEmCAYAAADWT9N8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x14eefc8d0>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "##PLOT CONFUSION MATRIX\n",
    "cnf_matrix = confusion_matrix(y_final_test, y_all3)\n",
    "np.set_printoptions(precision=2)\n",
    "class_names = [\"Right Hand\",\"Foot\"]\n",
    "\n",
    "# Plot non-normalized confusion matrix\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cnf_matrix, classes=class_names,title='Confusion matrix, without normalization')\n",
    "\n",
    "# Plot normalized confusion matrix\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cnf_matrix, classes=class_names, normalize=True,title='Normalized confusion matrix')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "1T-RJWhZmRzg"
   },
   "outputs": [],
   "source": [
    "#calculate score\n",
    "from sklearn.metrics import f1_score\n",
    "y_score = f1_score(y_final_test, y_all1, average=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "eLCLXHrzmRzi",
    "outputId": "212751b7-bf84-4b3a-f462-c9da11b65af7"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1.,  1.])"
      ]
     },
     "execution_count": 65,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "JM9cYU_lmRzk",
    "outputId": "faf399a8-d259-4017-ee89-a2d3d6e939d5"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEWCAYAAAB42tAoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x152dd4210>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Plot ROC curve\n",
    "\n",
    "count = 0\n",
    "iterator = 0\n",
    "real0 = []\n",
    "pred0 = []\n",
    "\n",
    "real1 = []\n",
    "pred1 = []\n",
    "\n",
    "\n",
    "real2 = []\n",
    "pred2 = []\n",
    "\n",
    "\n",
    "real3 = []\n",
    "pred3 = []\n",
    "\n",
    "while(iterator < len(y_final_test)):\n",
    "    if(y_final_test[iterator] == 0):\n",
    "        real0.append(0)\n",
    "        if(y_all1[iterator] == 0):\n",
    "            pred0.append(0)\n",
    "        else:\n",
    "            pred0.append(1)\n",
    "    else:\n",
    "        real0.append(1)\n",
    "        if(y_all1[iterator] == 0):\n",
    "            pred0.append(0)\n",
    "        else:\n",
    "            pred0.append(1)\n",
    "    iterator += 1\n",
    "    \n",
    "iterator = 0\n",
    "while(iterator < len(y_final_test)):\n",
    "    if(y_final_test[iterator] == 1):\n",
    "        real1.append(1)\n",
    "        if(y_all1[iterator] == 1):\n",
    "            pred1.append(1)\n",
    "        else:\n",
    "            pred1.append(0)\n",
    "    else:\n",
    "        real1.append(0)\n",
    "        if(y_all1[iterator] == 1):\n",
    "            pred1.append(1)\n",
    "        else:\n",
    "            pred1.append(0)\n",
    "    iterator += 1\n",
    "    \n",
    "    \n",
    "\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from itertools import cycle\n",
    "\n",
    "from sklearn import svm, datasets\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import label_binarize\n",
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "from scipy import interp\n",
    "\n",
    "fpr, tpr, _ = roc_curve(real0, pred0)\n",
    "roc_auc = auc(fpr, tpr)\n",
    "\n",
    "fpr1, tpr1, _ = roc_curve(real1, pred1)\n",
    "roc_auc1 = auc(fpr1, tpr1)\n",
    "\n",
    "plt.figure()\n",
    "lw = 2\n",
    "plt.plot(fpr, tpr, color='blue',\n",
    "         lw=lw, label='Right hand class ROC curve (area = %0.2f)' % roc_auc)\n",
    "\n",
    "plt.plot(fpr1, tpr1, color='green',\n",
    "         lw=lw, label='Right foot class ROC curve (area = %0.2f)' % roc_auc1)\n",
    "\n",
    "plt.plot([0, 1], [0, 1], color='darkorange', lw=lw, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('Receiver operating characteristic')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "u2BzK1qBmRzn",
    "outputId": "b12ca414-5d9a-414f-a922-046188a00c36"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEWCAYAAAB42tAoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQEoUIiISSIlCREQCKVGIiEig/w8QV+vNB0nkxgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x117ba4c10>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Plot precision recall curve\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from itertools import cycle\n",
    "\n",
    "from sklearn import svm, datasets\n",
    "from sklearn.metrics import precision_recall_curve\n",
    "from sklearn.metrics import average_precision_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import label_binarize\n",
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "\n",
    "\n",
    "# Compute Precision-Recall and plot curve\n",
    "precision = []\n",
    "recall = []\n",
    "average_precision = []\n",
    "\n",
    "precision, recall,thresholds = precision_recall_curve(real0,pred0)\n",
    "\n",
    "average_precision = average_precision_score(real0,pred0)\n",
    "\n",
    "\n",
    "precision1, recall1,thresholds1 = precision_recall_curve(real1,pred1)\n",
    "\n",
    "average_precision1 = average_precision_score(real1,pred1)\n",
    "\n",
    "\n",
    "# Plot Precision-Recall curve\n",
    "plt.clf()\n",
    "plt.plot(recall, precision, lw=lw, color='blue', label='Right hand class Precision-recall curve (area = {0:0.2f})'\n",
    "                   ''.format(average_precision))\n",
    "\n",
    "plt.plot(recall1, precision1, lw=lw, color='green', label='Right foot class Precision-recall curve (area = {0:0.2f})'\n",
    "                   ''.format(average_precision1))\n",
    "\n",
    "plt.xlabel('Recall')\n",
    "plt.ylabel('Precision')\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.title(\"Precision Recall Curve\")\n",
    "\n",
    "plt.legend(loc=\"lower left\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "G-bGpSRtmRzp"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "default_view": {},
   "name": "weighted_Subject av_dictionary based classification without BP and CSP_final.ipynb",
   "provenance": [],
   "version": "0.3.2",
   "views": {}
  },
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}