Word2Vec.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "# %load PCA.py\n",
    "%matplotlib inline\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import re\n",
    "import os\n",
    "import PyPDF2\n",
    "from nltk.stem.snowball import SnowballStemmer\n",
    "from nltk.corpus import stopwords\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfTransformer\n",
    "from sklearn.cluster import KMeans\n",
    "import matplotlib.pyplot as plt\n",
    "from mpl_toolkits.mplot3d import Axes3D\n",
    "from gensim.models.word2vec import Word2Vec\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "stemmer=SnowballStemmer(\"english\")\n",
    "stop = set(stopwords.words('english'))\n",
    "def pdf_to_text(pdf):\n",
    "\n",
    "    pdfReader = PyPDF2.PdfFileReader(pdf)\n",
    "    numPage=pdfReader.numPages\n",
    "    pdftext=\" \"\n",
    "    for page in range(numPage):\n",
    "        pageObj =pdfReader.getPage(page)\n",
    "        text= pageObj.extractText()\n",
    "        pagetext=\"\".join(text)\n",
    "        pdftext=\" \".join([pagetext, pdftext])\n",
    "    alltext = re.sub('[\\n]', '', pdftext)\n",
    "    all_words=alltext.split(\" \")\n",
    "    for i in range(len(all_words)):\n",
    "        all_words[i]= re.sub('[^A-Za-z0-9]+', '', all_words[i])\n",
    "    affterstemmer=[]\n",
    "    for word in all_words:\n",
    "        affterstemmer.append(stemmer.stem(word))\n",
    "    \n",
    "    afterstop=[]\n",
    "    for word in affterstemmer:\n",
    "        if word not in stop:\n",
    "            afterstop.append(word)\n",
    "    return \" \".join(afterstop)\n",
    "THIS_FOLDER = os.path.dirname(os.path.abspath(\"Word2Vec.ipynb\"))\n",
    "\n",
    "\n",
    "## Geting ready the training data\n",
    "data = pd.read_csv('resumes.csv')\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "PdfReadWarning: Xref table not zero-indexed. ID numbers for objects will be corrected. [pdf.py:1736]\n",
      "PdfReadWarning: Superfluous whitespace found in object header b'1' b'0' [pdf.py:1666]\n",
      "PdfReadWarning: Superfluous whitespace found in object header b'2' b'0' [pdf.py:1666]\n",
      "PdfReadWarning: Superfluous whitespace found in object header b'3' b'0' [pdf.py:1666]\n",
      "PdfReadWarning: Superfluous whitespace found in object header b'36' b'0' [pdf.py:1666]\n",
      "PdfReadWarning: Superfluous whitespace found in object header b'44' b'0' [pdf.py:1666]\n"
     ]
    }
   ],
   "source": [
    "X_train=[]\n",
    "for i in data['path']:\n",
    "    my_file = os.path.join(THIS_FOLDER, i)\n",
    "    pdf = open(my_file, 'rb')\n",
    "    X_train.append(pdf_to_text(pdf))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "word2vec = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "x_by_words=[]\n",
    "for i in X_train:\n",
    "    x_by_words.append(i.split())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "s = 1000\n",
    "\n",
    "model = Word2Vec(x_by_words, size=s, min_count=5, workers=10)\n",
    "w2v = {w: vec for w, vec in zip(model.wv.index2word, model.wv.syn0)}\n",
    "vocab = model.wv.index2word\n",
    "train_w2v = []\n",
    "for i in x_by_words:\n",
    "    ar =[0 for i in range(s)]\n",
    "    for j in i:\n",
    "        if j in vocab:\n",
    "            ar+=w2v[j]\n",
    "    train_w2v.append(ar)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(train_w2v, data['succes'], test_size = .001, random_state = 32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.cross_validation import cross_val_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.746811261976\n",
      "0.81581731617\n",
      "0.934285546347\n",
      "0.875036736613\n",
      "0.887527185094\n",
      "0.899988244284\n",
      "0.896843590196\n"
     ]
    }
   ],
   "source": [
    "clfs=[GaussianNB(),SGDClassifier(),LogisticRegression(),\n",
    "      DecisionTreeClassifier(),RandomForestClassifier(),\n",
    "      GradientBoostingClassifier(),AdaBoostClassifier()]\n",
    "for c in clfs:\n",
    "    clf = c.fit(x_train,y_train)\n",
    "    print(np.mean(cross_val_score(clf, train_w2v, data['succes'], cv=3)))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.92307692  0.96875     0.90625     0.96825397  0.92063492]\n",
      "0.937393162393\n"
     ]
    }
   ],
   "source": [
    "\n",
    "clf = LogisticRegression()\n",
    "cvs = cross_val_score(clf, x_train, y_train, cv=5)\n",
    "print(cvs)\n",
    "print(np.mean(cvs))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.824927996238\n",
      "0.943719508611\n",
      "0.909363427967\n",
      "0.549697290307\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.svm import SVC\n",
    "cvf_rbf= SVC(kernel = 'rbf', C= 1e3)\n",
    "cvf_linear= SVC(kernel = 'linear', C= 1e3)\n",
    "cvf_poly= SVC(kernel = 'poly', C= 1e3) \n",
    "cvf_sigmoid= SVC(kernel = 'sigmoid', C= 1e3) \n",
    "\n",
    "svc_clf= [cvf_rbf,cvf_linear,cvf_poly,cvf_sigmoid]\n",
    "\n",
    "for c in svc_clf:\n",
    "    clf = c.fit(x_train,y_train)\n",
    "    print(np.mean(cross_val_score(clf, train_w2v, data['succes'], cv=3)))\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.92307692  0.953125    0.921875    0.93650794  0.9047619 ]\n",
      "0.927869352869\n"
     ]
    }
   ],
   "source": [
    "clf = SVC(kernel = 'linear', C= 1e3)\n",
    "cvs = cross_val_score(clf, x_train, y_train, cv=5)\n",
    "print(cvs)\n",
    "print(np.mean(cvs))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid character in identifier (<ipython-input-23-6df232250e2f>, line 4)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;36m  File \u001b[0;32m\"<ipython-input-23-6df232250e2f>\"\u001b[0;36m, line \u001b[0;32m4\u001b[0m\n\u001b[0;31m    clf = GridSearchCV(svc,params,scoring=‘accuracy’)\u001b[0m\n\u001b[0m                                                   ^\u001b[0m\n\u001b[0;31mSyntaxError\u001b[0m\u001b[0;31m:\u001b[0m invalid character in identifier\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "svc = SVC(kernel = 'linear', C= 1e3)\n",
    "params = {'C': [1, 10, 100, 1000], 'kernel': ['linear']}\n",
    "clf = GridSearchCV(svc,params,scoring='accuracy')\n",
    "clf.fit(x_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(train_w2v, data['succes'], test_size = .001, random_state = 32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import learning_curve\n",
    "from sklearn.svm import SVC\n",
    "\n",
    "train_sizes, train_scores, test_scores = learning_curve(\n",
    "    SVC(kernel='linear'), x_train, y_train, train_sizes=[0.1 ,0.2 ,0.3 ,0.4 ,0.5 ,0.6 ,0.7 ,0.8 ,0.9 , 1.], cv=5)\n",
    "         \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEKCAYAAAAB0GKPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncycAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x125e020f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.xlabel(\"Training examples\")\n",
    "plt.ylabel(\"Score\")\n",
    "train_scores_mean = np.mean(train_scores, axis=1)\n",
    "train_scores_std = np.std(train_scores, axis=1)\n",
    "test_scores_mean = np.mean(test_scores, axis=1)\n",
    "test_scores_std = np.std(test_scores, axis=1)\n",
    "plt.grid()\n",
    "\n",
    "plt.fill_between(train_sizes, train_scores_mean - train_scores_std,\n",
    "                 train_scores_mean + train_scores_std, alpha=0.1,\n",
    "                 color=\"r\")\n",
    "plt.fill_between(train_sizes, test_scores_mean - test_scores_std,\n",
    "                 test_scores_mean + test_scores_std, alpha=0.1, color=\"g\")\n",
    "plt.plot(train_sizes, train_scores_mean, 'o-', color=\"r\",\n",
    "         label=\"Training score\")\n",
    "plt.plot(train_sizes, test_scores_mean, 'o-', color=\"g\",\n",
    "         label=\"Cross-validation score\")\n",
    "\n",
    "plt.legend(loc=\"best\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}