EEG_all_classifiers_from_raw_mne.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import warnings\n",
    "#warnings.filterwarnings('ignore')\n",
    "import pickle, os, glob\n",
    "from EEG.info import info_exp\n",
    "from EEG.converter import converter_mne\n",
    "import mne\n",
    "import numpy as np\n",
    "import sys"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "root = \"C:\\\\eeg\\\\01exp\\\\\"\n",
    "name = \"20161129_DBS_001\"\n",
    "\n",
    "def load(path,name):\n",
    "    with open(path + '/' + name, 'rb') as input:\n",
    "        data = pickle.load(input)\n",
    "        return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "test = load(root + name,'exp_data_aligned_start_150.cls')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=================================================================\n",
      "Experiment status:  True ; Num tests:  12\n",
      "Align status:  True\n",
      "Respondent name:  Ольга Ветрова\n",
      "Description: Эксперимент с обратной связью\n",
      "Labels_names: [ удалено, отдых, левая, правая ]\n",
      "Frequency: 1000\n",
      "Count train activities:  45\n",
      "Count tests activities: [ 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30 ]\n",
      "Time of activity: 4600\n",
      "Type_chanels: eeg  Montage: standard_1005\n",
      "------------------------------------------------------------------\n",
      "Chanels_names:\n",
      "------------------------------------------------------------------\n",
      "FC5        FT7        FC3        FCz        FC4        FT8       \n",
      "T3         C3         Cz         T4         C4        \n",
      "TP7        CP3        CPz        CP4        TP8       \n",
      "FC1        FC2        FC6        P4         C5        \n",
      "C1         C2         C6         P3         CP5       \n",
      "CP1        CP2        CP6        POz       \n",
      "------------------------------------------------------------------\n",
      "List tests\n",
      "------------------------------------------------------------------\n",
      "20161129_173025 20161129_173919 20161129_174505\n",
      "20161129_175011 20161129_175517\n",
      "20161129_180043 20161129_180632\n",
      "20161129_181557 20161129_182114\n",
      "20161129_182715 20161129_183224\n",
      "20161129_183740\n",
      "==================================================================\n"
     ]
    }
   ],
   "source": [
    "test.getinfo()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<RawArray  |  None, n_channels x n_times : 30 x 207000 (207.0 sec), ~47.4 MB, data loaded>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data  = converter_mne(test)\n",
    "epochs = data.train_epochs()\n",
    "raw = data.train_raw()\n",
    "raw.filter(10,16,phase='zero',filter_length='auto', fir_window='hamming',l_trans_bandwidth='auto', h_trans_bandwidth='auto')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsYAAABZCAYAAAA9x+IKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTzw+mDKOwHwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1d934780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Figure(720x108)\n"
     ]
    }
   ],
   "source": [
    "print epochs.plot_psd_topomap()\n",
    "picks = mne.pick_types(info=data.mne_info(), meg=False, eeg=True, misc=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbwAAAEsCAYAAAC41gacAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nPuH/A9o+5VHXGTbVAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1dbc4e48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Figure(432x288)\n"
     ]
    }
   ],
   "source": [
    "print epochs.plot_psd(area_mode='std', tmax=4.499,fmax=50.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'plt' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-18-663379f8ce60>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0mmne\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtime_frequency\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mpsd_multitaper\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mf\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0max\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msubplots\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m20\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m10\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m psds, freqs = psd_multitaper(raw, low_bias=True, tmax=4.499,\n\u001b[1;32m      4\u001b[0m                               \u001b[0mfmax\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m50.\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mproj\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mTrue\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpicks\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mpicks\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m                               n_jobs=1)\n",
      "\u001b[0;31mNameError\u001b[0m: name 'plt' is not defined"
     ]
    }
   ],
   "source": [
    "from mne.time_frequency import psd_multitaper\n",
    "f, ax = plt.subplots(figsize=(20, 10))\n",
    "psds, freqs = psd_multitaper(raw, low_bias=True, tmax=4.499,\n",
    "                              fmax=50., proj=True, picks=picks,\n",
    "                              n_jobs=1)\n",
    "psds = 10 * np.log10(psds)\n",
    "psds_mean = psds.mean(0)\n",
    "psds_std = psds.std(0)\n",
    "\n",
    "ax.grid(True)\n",
    "ax.plot(freqs, psds_mean, color='k')\n",
    "ax.fill_between(freqs, psds_mean - psds_std, psds_mean + psds_std,\n",
    "                color='k', alpha=.5)\n",
    "ax.set(title='Multitaper PSD', xlabel='Frequency',\n",
    "       ylabel='Power Spectral Density (dB)')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mne.set_log_level(verbose = \"ERROR\")\n",
    "# get test list\n",
    "test_list = []\n",
    "for i in range(0,data.num_tests):\n",
    "    epochs_test = data.test_epochs(i)\n",
    "    Y_test = epochs_test.events[:,-1]\n",
    "    X_test = epochs_test.get_data()\n",
    "    test_list.append({'Y_test':Y_test,'X_test':X_test})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import scipy.signal as spsig\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def data_prep(data, sfreq, fmin, fmax):\n",
    "    [b_high, a_high] = spsig.butter(4, float(fmin) / (sfreq / 2), 'high')\n",
    "    [b_low, a_low] = spsig.butter(4, float(fmax) / (sfreq / 2), 'low')\n",
    "\n",
    "    chunk_high = spsig.lfilter(b_high, a_high, data)\n",
    "    chunk_low = spsig.lfilter(b_low, a_low, chunk_high)\n",
    "    return chunk_low\n",
    "\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\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')\n",
    "    pass\n",
    "\n",
    "def plot_mult_conf_matrices(conf_matrices,classes,method='mean', mode = 'plot', normalize = True, title = 'Confusion matrix'):\n",
    "    if(len(conf_matrices) < 2):\n",
    "        print \"is not a matrices set\"\n",
    "        return 0\n",
    "    if(conf_matrices[0].shape[0] != conf_matrices[0].shape[1]):\n",
    "        print \"is not a square matrix\"\n",
    "        return 0\n",
    "\n",
    "    n = conf_matrices[0].shape[0]\n",
    "    \n",
    "    if(method == 'mean'):\n",
    "        matrices = np.mean([confusion_matrix[n-j] for j in range(1,n)], axis = 0)\n",
    "    elif(method == 'max'):\n",
    "        matrices = np.max([confusion_matrix[n-j] for j in range(1,n)], axis = 0)\n",
    "    elif(method == 'min'):\n",
    "        matrices = np.min([confusion_matrix[n-j] for j in range(1,n)], axis = 0)\n",
    "    else:\n",
    "        print \"incorrect method\";\n",
    "    if(mode == 'plot'):\n",
    "        plt.figure()\n",
    "        plot_confusion_matrix(matrices, classes,\n",
    "                      title=title, normalize = normalize)\n",
    "    else:\n",
    "        return matrices"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LDA Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.metrics import accuracy_score\n",
    "from mne.decoding import CSP\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn import metrics\n",
    "import itertools\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = LinearDiscriminantAnalysis()\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Naive Bayes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\User\\Anaconda2\\lib\\site-packages\\sklearn\\metrics\\classification.py:1113: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 1.0\n",
      "MEAN: 0.802777777778\n",
      "MIN: 0.533333333333\n",
      "DISP: 0.136394263302\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       1.00      1.00      1.00        10\n",
      "          2       0.77      1.00      0.87        10\n",
      "          3       1.00      0.70      0.82        10\n",
      "\n",
      "avg / total       0.92      0.90      0.90        30\n",
      "\n",
      "Normalized confusion matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUkAAAEmCAYAAADvKGInAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x4d7b7cf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.naive_bayes import GaussianNB\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = GaussianNB()\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## K-nearest neighbours"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 0.966666666667\n",
      "MEAN: 0.866666666667\n",
      "MIN: 0.766666666667\n",
      "DISP: 0.0561083607687\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       0.91      1.00      0.95        10\n",
      "          2       0.82      0.90      0.86        10\n",
      "          3       1.00      0.80      0.89        10\n",
      "\n",
      "avg / total       0.91      0.90      0.90        30\n",
      "\n",
      "Normalized confusion matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUkAAAEmCAYAAADvKGInAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nB79O6FzFsOcMAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x4d2d45c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = KNeighborsClassifier()\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classification and Regression Trees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 0.933333333333\n",
      "MEAN: 0.786111111111\n",
      "MIN: 0.6\n",
      "DISP: 0.0957024065386\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       0.67      1.00      0.80        10\n",
      "          2       0.78      0.70      0.74        10\n",
      "          3       1.00      0.60      0.75        10\n",
      "\n",
      "avg / total       0.81      0.77      0.76        30\n",
      "\n",
      "Normalized confusion matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUkAAAEmCAYAAADvKGInAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x4d3bb198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = DecisionTreeClassifier()\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SVM (Support Vector Machines)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 1.0\n",
      "MEAN: 0.844444444444\n",
      "MIN: 0.666666666667\n",
      "DISP: 0.109994388185\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       1.00      1.00      1.00        10\n",
      "          2       0.91      1.00      0.95        10\n",
      "          3       1.00      0.90      0.95        10\n",
      "\n",
      "avg / total       0.97      0.97      0.97        30\n",
      "\n",
      "Normalized confusion matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUkAAAEmCAYAAADvKGInAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkjQzy8NJ0swsDydJM7M8/j+/2iiGXI/GgQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1d2e5160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.svm import SVC\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = SVC()\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 0.933333333333\n",
      "MEAN: 0.783333333333\n",
      "MIN: 0.6\n",
      "DISP: 0.102288625775\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       1.00      1.00      1.00        10\n",
      "          2       0.71      1.00      0.83        10\n",
      "          3       1.00      0.60      0.75        10\n",
      "\n",
      "avg / total       0.90      0.87      0.86        30\n",
      "\n",
      "Normalized confusion matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUMAAAEmCAYAAAD4CvLuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niogASoYiIoCSoYgIoGQoIgIoGYqIAPD/BNx5Xx6UQrkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1d7dbb38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.neural_network import MLPClassifier\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = MLPClassifier(solver='lbfgs', alpha=1e-5, hidden_layer_sizes=(8, 15), random_state=1)\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 0.966666666667\n",
      "MEAN: 0.727777777778\n",
      "MIN: 0.6\n",
      "DISP: 0.112079116677\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       1.00      1.00      1.00        10\n",
      "          2       0.62      1.00      0.77        10\n",
      "          3       1.00      0.40      0.57        10\n",
      "\n",
      "avg / total       0.88      0.80      0.78        30\n",
      "\n",
      "Normalized confusion matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUkAAAEmCAYAAADvKGInAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nHIxGcGnqAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1defd438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = LogisticRegression()\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 0.966666666667\n",
      "MEAN: 0.775\n",
      "MIN: 0.366666666667\n",
      "DISP: 0.165061716853\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       1.00      1.00      1.00        10\n",
      "          2       0.77      1.00      0.87        10\n",
      "          3       1.00      0.70      0.82        10\n",
      "\n",
      "avg / total       0.92      0.90      0.90        30\n",
      "\n",
      "Normalized confusion matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUkAAAEmCAYAAADvKGInAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1d59eef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from skrvm import RVC\n",
    "\n",
    "Y_train = epochs.events[:,-1]\n",
    "X_train = epochs.get_data()\n",
    "X_train = data_prep(X_train, 1000, 6, 16)\n",
    "\n",
    "# train classifier\n",
    "svc = RVC(alpha=1e-06, beta=1e-06, beta_fixed=False, bias_used=True, coef0=0.0,\n",
    "coef1=None, degree=3, kernel='rbf', n_iter=3000, n_iter_posterior=50,\n",
    "threshold_alpha=1000000000.0, tol=0.001, verbose=False)\n",
    "csp = CSP(n_components=4, reg='ledoit_wolf', log=True)\n",
    "\n",
    "\n",
    "X_train = csp.fit_transform(X_train, Y_train)\n",
    "svc.fit(X_train, Y_train)\n",
    "\n",
    "scores = []\n",
    "classification_report = []\n",
    "confusion_matrix = []\n",
    "for i in range(0, len(test_list)):\n",
    "    X_test = data_prep(test_list[i]['X_test'], 1000, 6, 16)\n",
    "    X_test = csp.transform(X_test)\n",
    "    scores.append(svc.score(X_test, test_list[i]['Y_test']))\n",
    "    classification_report.append(metrics.classification_report(test_list[i]['Y_test'], svc.predict(X_test)))\n",
    "    confusion_matrix.append(metrics.confusion_matrix(test_list[i]['Y_test'],svc.predict(X_test)))\n",
    "\n",
    "print \"MAX:\", np.max(scores)\n",
    "print \"MEAN:\", np.mean(scores)\n",
    "print \"MIN:\", np.min(scores)\n",
    "print \"DISP:\", np.std(scores)\n",
    "print classification_report[0]\n",
    "\n",
    "plot_mult_conf_matrices(confusion_matrix,[u'Rest',u'left',u'right'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": 2
}