Data Pre-processing for essay grader.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import nltk\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "import keras as k\n",
    "from keras.preprocessing.sequence import pad_sequences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data=pd.read_csv('training_set_rel3.tsv',encoding='ISO-8859-1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data=data[data['essay_set']<7]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data=data.reset_index()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Vocabulary "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "vocab=pd.read_csv('glove.6B/vocab.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "vocab=vocab.drop(vocab.columns[0],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "vocab=vocab.drop(vocab.columns[1],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "vocab.index=vocab.index+1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dictionary "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "word_id={}\n",
    "for i in range(len(vocab)):\n",
    "    word_id[vocab.loc[i+1,'word']]=i+1\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from nltk.stem import WordNetLemmatizer\n",
    "from nltk.stem import PorterStemmer\n",
    "lemmatizer = WordNetLemmatizer()\n",
    "#stemming=PorterStemmer()\n",
    "\n",
    "            \n",
    "def doc_to_id(h):\n",
    "    l=[]\n",
    "    m=[]\n",
    "    a=nltk.tokenize.word_tokenize(h)\n",
    "    for w in a:\n",
    "        try:\n",
    "            n1=lemmatizer.lemmatize(w.lower())\n",
    "            #n=stemming.stem(n1)\n",
    "            l.append(word_id[n1])\n",
    "        except:\n",
    "                m.append(w.lower())\n",
    "    return l,m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "doc_to_id('my name is hardik')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Essay to Id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data1=pd.DataFrame(data['essay'])\n",
    "data1['domain_score']=data['domain1_score']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "word_id1=[]\n",
    "wrong_word=[]\n",
    "for i in range(len(data1)):\n",
    "    p,q=doc_to_id(data1['essay'][i])\n",
    "    word_id1.append(p)\n",
    "    wrong_word.append(q)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "seq_length=[len(word_id1[i]) for i in range(len(word_id1))]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  Study of sequence length of essays"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 269,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fbfbd898410>"
      ]
     },
     "execution_count": 269,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAigAAAFoCAYAAAB9vQB0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbfc2695490>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(seq_length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data1['seq_length']=seq_length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#Removing essays who have length greater than 550\n",
    "data1=data1[data1['seq_length']<550]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data1=data1.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "word_id1=[]\n",
    "wrong_word=[]\n",
    "for i in range(len(data1)):\n",
    "    p,q=doc_to_id(data1['essay'][i])\n",
    "    word_id1.append(p)\n",
    "    wrong_word.append(q)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "seq_length=[len(word_id1[i]) for i in range(len(word_id1))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fbfc12fc550>"
      ]
     },
     "execution_count": 276,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiQAAAF9CAYAAAA0mnPSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbfc0f79a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(data1['seq_length'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data1.drop(data.index[8936],inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Removing other columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data1=data1[~data1['domain_score'].isnull()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data1=data1.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "word_id1=[]\n",
    "wrong_word=[]\n",
    "for i in range(len(data1)):\n",
    "    p,q=doc_to_id(data1['essay'][i])\n",
    "    word_id1.append(p)\n",
    "    wrong_word.append(q)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "seq_length=[len(word_id1[i]) for i in range(len(word_id1))]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Padding for Seq2Seq model and make suitable output for softmax layer "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input1=pad_sequences(word_id1,maxlen=550,padding='post')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "output=k.utils.np_utils.to_categorical(data1['domain_score'],nb_classes=13)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Save Final Data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "np.save('Essay grader/input',input1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 303,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.save('Essay grader/output',output)"
   ]
  }
 ],
 "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": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}