svm_baseline.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import print_function\n",
    "from IPython.display import display, HTML\n",
    "import os\n",
    "import sys\n",
    "import csv\n",
    "import time\n",
    "import statistics\n",
    "import numpy as np\n",
    "import string\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import KFold, train_test_split\n",
    "from sklearn import svm\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import average_precision_score, accuracy_score, recall_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Chin data shape\n",
      "(804, 79)\n",
      "(804, 76)\n",
      "(804, 1)\n",
      "Eng data shape\n",
      "(3570, 79)\n",
      "(3570, 76)\n",
      "(3570, 1)\n"
     ]
    }
   ],
   "source": [
    "#Read in the csv\n",
    "#Get features columns and convert it to numpy array\n",
    "chinDataDf = pd.read_csv(\"data_with_features/chin_cleaned_data_f.csv\")\n",
    "chinFeaturesDf = chinDataDf.drop(['depressed', chinDataDf.columns[0],'text'], axis=1)\n",
    "chinFeaturesNpArray = chinFeaturesDf.as_matrix()\n",
    "chinFeaturesNpArray[np.isnan(chinFeaturesNpArray)] = 0.0\n",
    "#Get label columns and convert it to numpy array\n",
    "chinLabelsDf = chinDataDf[['depressed']]\n",
    "chinLabelsNpArray = chinLabelsDf.as_matrix()\n",
    "#convert all -1 to 0\n",
    "chinLabelsNpArray[chinLabelsNpArray==-1] = 0\n",
    "print(\"Chin data shape\")\n",
    "print(chinDataDf.shape)\n",
    "print(chinFeaturesDf.shape)\n",
    "print(chinLabelsDf.shape)\n",
    "engDataDf = pd.read_csv(\"data_with_features/eng_cleaned_data_f.csv\")\n",
    "engFeaturesDf = engDataDf.drop(['depressed', engDataDf.columns[0],'text'], axis=1)\n",
    "engFeaturesNpArray = engFeaturesDf.as_matrix()\n",
    "engFeaturesNpArray[np.isnan(engFeaturesNpArray)] = 0.0\n",
    "#Get label columns and convert it to numpy array\n",
    "engLabelsDf = engDataDf[['depressed']]\n",
    "engLabelsNpArray = engLabelsDf.as_matrix()\n",
    "#convert all -1 to 0\n",
    "engLabelsNpArray[engLabelsNpArray==-1] = 0\n",
    "print(\"Eng data shape\")\n",
    "print(engDataDf.shape)\n",
    "print(engFeaturesDf.shape)\n",
    "print(engLabelsDf.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Set parameters for grid search \n",
    "tuned_parameters = [{'kernel': ['rbf'], 'gamma': [1e-3, 1e-4],\n",
    "                     'C': [0.1, 1, 10, 100, 1000]},\n",
    "                    {'kernel': ['linear'], 'C': [0.1, 1, 10, 100, 1000]}]\n",
    "scores = ['accuracy', 'precision', 'recall']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SVM baseline for Chinese data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Train:Test ratio = 80%:20%\n",
    "#Split between training set and testing set, perform k-fold cross validation on testing set after training\n",
    "chinXTrain, chinXTest, chinYTrain, chinYTest = train_test_split(chinFeaturesNpArray, chinLabelsNpArray, test_size=0.2, shuffle=True, random_state=228, stratify=chinLabelsNpArray)\n",
    "r1, c1 = chinYTrain.shape\n",
    "r2, c2 = chinYTest.shape\n",
    "chinYTrain = chinYTrain.reshape(r1,)\n",
    "chinYTest = chinYTest.reshape(r2,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Tuning hyper-parameters for accuracy\n",
      "\n",
      "Best parameters set found on development set:\n",
      "\n",
      "SVC(C=10, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "  tol=0.001, verbose=False)\n",
      "\n",
      "Grid scores on development set:\n",
      "\n",
      "0.659 (+/-0.001) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.001}\n",
      "0.659 (+/-0.001) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.0001}\n",
      "0.670 (+/-0.010) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.001}\n",
      "0.661 (+/-0.003) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.0001}\n",
      "0.664 (+/-0.011) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.001}\n",
      "0.664 (+/-0.006) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.0001}\n",
      "0.658 (+/-0.005) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.001}\n",
      "0.669 (+/-0.008) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.0001}\n",
      "0.714 (+/-0.017) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.001}\n",
      "0.664 (+/-0.009) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.0001}\n",
      "0.669 (+/-0.007) for {'kernel': 'linear', 'C': 0.1}\n",
      "0.700 (+/-0.020) for {'kernel': 'linear', 'C': 1}\n",
      "0.771 (+/-0.023) for {'kernel': 'linear', 'C': 10}\n",
      "0.759 (+/-0.013) for {'kernel': 'linear', 'C': 100}\n",
      "0.756 (+/-0.015) for {'kernel': 'linear', 'C': 1000}\n",
      "\n",
      "Detailed classification report:\n",
      "\n",
      "The model is trained on the full development set.\n",
      "The scores are computed on the full evaluation set.\n",
      "\n",
      "Average precision-recall score: 0.52\n",
      "Recall score: 0.51\n",
      "Accuracy: 0.751552795031\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          0       0.78      0.88      0.82       106\n",
      "          1       0.68      0.51      0.58        55\n",
      "\n",
      "avg / total       0.74      0.75      0.74       161\n",
      "\n",
      "\n",
      "# Tuning hyper-parameters for precision\n",
      "\n",
      "Best parameters set found on development set:\n",
      "\n",
      "SVC(C=10, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "  tol=0.001, verbose=False)\n",
      "\n",
      "Grid scores on development set:\n",
      "\n",
      "0.000 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.001}\n",
      "0.000 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.0001}\n",
      "0.661 (+/-0.140) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.001}\n",
      "0.581 (+/-0.191) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.0001}\n",
      "0.594 (+/-0.139) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.001}\n",
      "0.611 (+/-0.117) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.0001}\n",
      "0.547 (+/-0.118) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.001}\n",
      "0.663 (+/-0.112) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.0001}\n",
      "0.637 (+/-0.039) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.001}\n",
      "0.546 (+/-0.088) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.0001}\n",
      "0.681 (+/-0.132) for {'kernel': 'linear', 'C': 0.1}\n",
      "0.705 (+/-0.092) for {'kernel': 'linear', 'C': 1}\n",
      "0.715 (+/-0.044) for {'kernel': 'linear', 'C': 10}\n",
      "0.653 (+/-0.016) for {'kernel': 'linear', 'C': 100}\n",
      "0.643 (+/-0.017) for {'kernel': 'linear', 'C': 1000}\n",
      "\n",
      "Detailed classification report:\n",
      "\n",
      "The model is trained on the full development set.\n",
      "The scores are computed on the full evaluation set.\n",
      "\n",
      "Average precision-recall score: 0.52\n",
      "Recall score: 0.51\n",
      "Accuracy: 0.751552795031\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          0       0.78      0.88      0.82       106\n",
      "          1       0.68      0.51      0.58        55\n",
      "\n",
      "avg / total       0.74      0.75      0.74       161\n",
      "\n",
      "\n",
      "# Tuning hyper-parameters for recall\n",
      "\n",
      "Best parameters set found on development set:\n",
      "\n",
      "SVC(C=1000, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "  tol=0.001, verbose=False)\n",
      "\n",
      "Grid scores on development set:\n",
      "\n",
      "0.000 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.001}\n",
      "0.000 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.0001}\n",
      "0.068 (+/-0.018) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.001}\n",
      "0.023 (+/-0.007) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.0001}\n",
      "0.073 (+/-0.017) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.001}\n",
      "0.050 (+/-0.008) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.0001}\n",
      "0.073 (+/-0.017) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.001}\n",
      "0.073 (+/-0.011) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.0001}\n",
      "0.365 (+/-0.041) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.001}\n",
      "0.096 (+/-0.021) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.0001}\n",
      "0.055 (+/-0.012) for {'kernel': 'linear', 'C': 0.1}\n",
      "0.224 (+/-0.024) for {'kernel': 'linear', 'C': 1}\n",
      "0.552 (+/-0.032) for {'kernel': 'linear', 'C': 10}\n",
      "0.621 (+/-0.039) for {'kernel': 'linear', 'C': 100}\n",
      "0.630 (+/-0.040) for {'kernel': 'linear', 'C': 1000}\n",
      "\n",
      "Detailed classification report:\n",
      "\n",
      "The model is trained on the full development set.\n",
      "The scores are computed on the full evaluation set.\n",
      "\n",
      "Average precision-recall score: 0.52\n",
      "Recall score: 0.60\n",
      "Accuracy: 0.745341614907\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "          0       0.80      0.82      0.81       106\n",
      "          1       0.63      0.60      0.62        55\n",
      "\n",
      "avg / total       0.74      0.75      0.74       161\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for score in scores:\n",
    "    print(\"# Tuning hyper-parameters for %s\" % score)\n",
    "    print()\n",
    "\n",
    "    clf = GridSearchCV(svm.SVC(C=1), tuned_parameters, cv=5, scoring=score)\n",
    "    clf.fit(chinXTrain, chinYTrain)\n",
    "\n",
    "    print(\"Best parameters set found on development set:\")\n",
    "    print()\n",
    "    print(clf.best_estimator_)\n",
    "    print()\n",
    "    print(\"Grid scores on development set:\")\n",
    "    print()\n",
    "    for params, meanScore, scores in clf.grid_scores_:\n",
    "        print(\"%0.3f (+/-%0.03f) for %r\"\n",
    "              % (meanScore, scores.std() / 2, params))\n",
    "    print()\n",
    "\n",
    "    print(\"Detailed classification report:\")\n",
    "    print()\n",
    "    print(\"The model is trained on the full development set.\")\n",
    "    print(\"The scores are computed on the full evaluation set.\")\n",
    "    print()\n",
    "    yTrue, yPred = chinYTest, clf.best_estimator_.predict(chinXTest)\n",
    "    svm_precision_score = average_precision_score(yTrue, yPred)\n",
    "    svm_recall_score = recall_score(yTrue, yPred, average='binary')\n",
    "    print('Average precision-recall score: {0:0.2f}'.format(svm_precision_score))\n",
    "    print('Recall score: {0:0.2f}'.format(svm_recall_score))\n",
    "    print('Accuracy: ' + str(accuracy_score(yTrue, yPred)))\n",
    "    print(classification_report(yTrue, yPred))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SVM baseline for English data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Set parameters for grid search \n",
    "tuned_parameters = [{'kernel': ['rbf'], 'gamma': [1e-3, 1e-4],\n",
    "                     'C': [0.1, 1, 10, 100, 1000]},\n",
    "                    {'kernel': ['linear'], 'C': [0.1, 1, 10, 100, 1000]}]\n",
    "engScores = ['accuracy', 'precision', 'recall']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Train:Test ratio = 80%:20%\n",
    "#Split between training set and testing set, perform k-fold cross validation on testing set after training\n",
    "engXTrain, engXTest, engYTrain, engYTest = train_test_split(engFeaturesNpArray, engLabelsNpArray, test_size=0.2, shuffle=True, random_state=228, stratify=engLabelsNpArray)\n",
    "r1, c1 = engYTrain.shape\n",
    "r2, c2 = engYTest.shape\n",
    "engYTrain = engYTrain.reshape(r1,)\n",
    "engYTest = engYTest.reshape(r2,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Tuning hyper-parameters for accuracy\n",
      "\n",
      "Best parameters set found on development set:\n",
      "\n",
      "SVC(C=1000, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "  tol=0.001, verbose=False)\n",
      "\n",
      "Grid scores on development set:\n",
      "\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.0001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.0001}\n",
      "0.707 (+/-0.009) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.0001}\n",
      "0.800 (+/-0.009) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.001}\n",
      "0.707 (+/-0.009) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.0001}\n",
      "0.812 (+/-0.011) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.001}\n",
      "0.800 (+/-0.009) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.0001}\n",
      "0.791 (+/-0.010) for {'kernel': 'linear', 'C': 0.1}\n",
      "0.808 (+/-0.011) for {'kernel': 'linear', 'C': 1}\n",
      "0.812 (+/-0.010) for {'kernel': 'linear', 'C': 10}\n",
      "0.822 (+/-0.009) for {'kernel': 'linear', 'C': 100}\n",
      "0.824 (+/-0.008) for {'kernel': 'linear', 'C': 1000}\n",
      "\n",
      "Detailed classification report:\n",
      "\n",
      "The model is trained on the full development set.\n",
      "The scores are computed on the full evaluation set.\n",
      "\n",
      "Accuracy: 0.843137254902\n",
      "Average precision-recall score: 0.80\n",
      "Recall score: 0.86\n",
      "\n",
      "# Tuning hyper-parameters for precision\n",
      "\n",
      "Best parameters set found on development set:\n",
      "\n",
      "SVC(C=1000, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "  tol=0.001, verbose=False)\n",
      "\n",
      "Grid scores on development set:\n",
      "\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.0001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.0001}\n",
      "0.656 (+/-0.007) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.001}\n",
      "0.529 (+/-0.000) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.0001}\n",
      "0.815 (+/-0.012) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.001}\n",
      "0.656 (+/-0.007) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.0001}\n",
      "0.807 (+/-0.012) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.001}\n",
      "0.815 (+/-0.012) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.0001}\n",
      "0.818 (+/-0.015) for {'kernel': 'linear', 'C': 0.1}\n",
      "0.806 (+/-0.012) for {'kernel': 'linear', 'C': 1}\n",
      "0.808 (+/-0.013) for {'kernel': 'linear', 'C': 10}\n",
      "0.819 (+/-0.011) for {'kernel': 'linear', 'C': 100}\n",
      "0.820 (+/-0.008) for {'kernel': 'linear', 'C': 1000}\n",
      "\n",
      "Detailed classification report:\n",
      "\n",
      "The model is trained on the full development set.\n",
      "The scores are computed on the full evaluation set.\n",
      "\n",
      "Accuracy: 0.843137254902\n",
      "Average precision-recall score: 0.80\n",
      "Recall score: 0.86\n",
      "\n",
      "# Tuning hyper-parameters for recall\n",
      "\n",
      "Best parameters set found on development set:\n",
      "\n",
      "SVC(C=0.1, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma=0.001, kernel='rbf',\n",
      "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "  tol=0.001, verbose=False)\n",
      "\n",
      "Grid scores on development set:\n",
      "\n",
      "1.000 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.001}\n",
      "1.000 (+/-0.000) for {'kernel': 'rbf', 'C': 0.1, 'gamma': 0.0001}\n",
      "1.000 (+/-0.000) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.001}\n",
      "1.000 (+/-0.000) for {'kernel': 'rbf', 'C': 1, 'gamma': 0.0001}\n",
      "0.937 (+/-0.008) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.001}\n",
      "1.000 (+/-0.000) for {'kernel': 'rbf', 'C': 10, 'gamma': 0.0001}\n",
      "0.807 (+/-0.013) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.001}\n",
      "0.937 (+/-0.008) for {'kernel': 'rbf', 'C': 100, 'gamma': 0.0001}\n",
      "0.849 (+/-0.011) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.001}\n",
      "0.807 (+/-0.013) for {'kernel': 'rbf', 'C': 1000, 'gamma': 0.0001}\n",
      "0.780 (+/-0.009) for {'kernel': 'linear', 'C': 0.1}\n",
      "0.840 (+/-0.014) for {'kernel': 'linear', 'C': 1}\n",
      "0.849 (+/-0.008) for {'kernel': 'linear', 'C': 10}\n",
      "0.853 (+/-0.009) for {'kernel': 'linear', 'C': 100}\n",
      "0.855 (+/-0.010) for {'kernel': 'linear', 'C': 1000}\n",
      "\n",
      "Detailed classification report:\n",
      "\n",
      "The model is trained on the full development set.\n",
      "The scores are computed on the full evaluation set.\n",
      "\n",
      "Accuracy: 0.529411764706\n",
      "Average precision-recall score: 0.53\n",
      "Recall score: 1.00\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for score in engScores:\n",
    "    print(\"# Tuning hyper-parameters for %s\" % score)\n",
    "    print()\n",
    "\n",
    "    clf = GridSearchCV(svm.SVC(C=1), tuned_parameters, cv=5, scoring=score)\n",
    "    clf.fit(engXTrain, engYTrain)\n",
    "\n",
    "    print(\"Best parameters set found on development set:\")\n",
    "    print()\n",
    "    print(clf.best_estimator_)\n",
    "    print()\n",
    "    print(\"Grid scores on development set:\")\n",
    "    print()\n",
    "    for params, meanScore, scores in clf.grid_scores_:\n",
    "        print(\"%0.3f (+/-%0.03f) for %r\"\n",
    "              % (meanScore, scores.std() / 2, params))\n",
    "    print()\n",
    "\n",
    "    print(\"Detailed classification report:\")\n",
    "    print()\n",
    "    print(\"The model is trained on the full development set.\")\n",
    "    print(\"The scores are computed on the full evaluation set.\")\n",
    "    print()\n",
    "    yTrue, yPred = engYTest, clf.best_estimator_.predict(engXTest)\n",
    "    svm_precision_score = average_precision_score(yTrue, yPred)\n",
    "    svm_recall_score = recall_score(yTrue, yPred, average='binary')\n",
    "    print('Accuracy: ' + str(accuracy_score(yTrue, yPred)))\n",
    "    print('Average precision-recall score: {0:0.2f}'.format(svm_precision_score))\n",
    "    print('Recall score: {0:0.2f}'.format(svm_recall_score))\n",
    "    #yTrue, yPred = engYTest, clf.predict(engXTest)\n",
    "    #print(classification_report(yTrue, yPred))\n",
    "    print()"
   ]
  },
  {
   "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": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}