Assignment3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\andy\\Anaconda3\\lib\\site-packages\\sklearn\\cross_validation.py:41: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "import pickle,pprint\n",
    "\n",
    "import sqlite3\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import nltk\n",
    "\n",
    "\n",
    "import re\n",
    "\n",
    "\n",
    "from nltk.corpus import stopwords\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from gensim.models import Word2Vec\n",
    "\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn import cross_validation\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.cross_validation import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pkl_file = open('data.pkl', 'rb')\n",
    "filtered_data = pickle.load(pkl_file)\n",
    "#pprint.pprint(filtered_data)\n",
    "pkl_file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['ProductId', 'UserId', 'ProfileName', 'HelpfulnessNumerator',\n",
       "       'HelpfulnessDenominator', 'Time', 'Summary', 'Text', 'Sentiment',\n",
       "       'CleanedText'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_data.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Filtering 2000 Positive and 2000 Negaitive review since more data points are creating :MemoryError"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_data.sort_values('Time',inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_10k=filtered_data.iloc[:10000,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1130     961718400\n",
       "1129     962236800\n",
       "7338    1075420800\n",
       "9545    1081209600\n",
       "3439    1081555200\n",
       "6709    1083456000\n",
       "1097    1088467200\n",
       "9542    1089504000\n",
       "5818    1090972800\n",
       "5819    1091059200\n",
       "1160    1091318400\n",
       "1157    1092787200\n",
       "9952    1094860800\n",
       "5068    1100563200\n",
       "1177    1106438400\n",
       "10      1107820800\n",
       "6681    1108080000\n",
       "2457    1108771200\n",
       "8029    1109289600\n",
       "6805    1113782400\n",
       "Name: Time, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_10k.Time.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_10k.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "label=filtered_10k.Sentiment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_10k.drop('Sentiment',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>index</th>\n",
       "      <th>ProductId</th>\n",
       "      <th>UserId</th>\n",
       "      <th>ProfileName</th>\n",
       "      <th>HelpfulnessNumerator</th>\n",
       "      <th>HelpfulnessDenominator</th>\n",
       "      <th>Time</th>\n",
       "      <th>Summary</th>\n",
       "      <th>Text</th>\n",
       "      <th>CleanedText</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1130</td>\n",
       "      <td>B00002Z754</td>\n",
       "      <td>A29Z5PI9BW2PU3</td>\n",
       "      <td>Robbie</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>961718400</td>\n",
       "      <td>Great Product</td>\n",
       "      <td>This was a really good idea and the final prod...</td>\n",
       "      <td>realli good idea final product outstand use de...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1129</td>\n",
       "      <td>B00002Z754</td>\n",
       "      <td>A3B8RCEI0FXFI6</td>\n",
       "      <td>B G Chase</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>962236800</td>\n",
       "      <td>WOW Make your own 'slickers' !</td>\n",
       "      <td>I just received my shipment and could hardly w...</td>\n",
       "      <td>receiv shipment could hard wait tri product lo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7338</td>\n",
       "      <td>B0000EIE2Z</td>\n",
       "      <td>A3M174IC0VXOS2</td>\n",
       "      <td>Gail Cooke</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>1075420800</td>\n",
       "      <td>BEST BLUEBERRIES</td>\n",
       "      <td>In the winter when fresh blueberries exceed ou...</td>\n",
       "      <td>winter fresh blueberri exceed food budget dri ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9545</td>\n",
       "      <td>B0000VYKXC</td>\n",
       "      <td>A3L88AH6MABEDL</td>\n",
       "      <td>Javier Denison</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1081209600</td>\n",
       "      <td>Best coffee I've tasted. I try others but alwa...</td>\n",
       "      <td>If you like a strong but smooth flavorful coff...</td>\n",
       "      <td>like strong smooth flavor coffe tri other keep...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3439</td>\n",
       "      <td>B00016UX0K</td>\n",
       "      <td>AF1PV3DIC0XM7</td>\n",
       "      <td>Robert Ashton</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1081555200</td>\n",
       "      <td>Classic Condiment</td>\n",
       "      <td>Mae Ploy Sweet Chili Sauce is becoming a stand...</td>\n",
       "      <td>mae ploy sweet chili sauc becom standard condi...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   index   ProductId          UserId     ProfileName  HelpfulnessNumerator  \\\n",
       "0   1130  B00002Z754  A29Z5PI9BW2PU3          Robbie                     7   \n",
       "1   1129  B00002Z754  A3B8RCEI0FXFI6       B G Chase                    10   \n",
       "2   7338  B0000EIE2Z  A3M174IC0VXOS2      Gail Cooke                     3   \n",
       "3   9545  B0000VYKXC  A3L88AH6MABEDL  Javier Denison                     0   \n",
       "4   3439  B00016UX0K   AF1PV3DIC0XM7   Robert Ashton                     1   \n",
       "\n",
       "   HelpfulnessDenominator        Time  \\\n",
       "0                       7   961718400   \n",
       "1                      10   962236800   \n",
       "2                       3  1075420800   \n",
       "3                       0  1081209600   \n",
       "4                       2  1081555200   \n",
       "\n",
       "                                             Summary  \\\n",
       "0                                      Great Product   \n",
       "1                     WOW Make your own 'slickers' !   \n",
       "2                                   BEST BLUEBERRIES   \n",
       "3  Best coffee I've tasted. I try others but alwa...   \n",
       "4                                  Classic Condiment   \n",
       "\n",
       "                                                Text  \\\n",
       "0  This was a really good idea and the final prod...   \n",
       "1  I just received my shipment and could hardly w...   \n",
       "2  In the winter when fresh blueberries exceed ou...   \n",
       "3  If you like a strong but smooth flavorful coff...   \n",
       "4  Mae Ploy Sweet Chili Sauce is becoming a stand...   \n",
       "\n",
       "                                         CleanedText  \n",
       "0  realli good idea final product outstand use de...  \n",
       "1  receiv shipment could hard wait tri product lo...  \n",
       "2  winter fresh blueberri exceed food budget dri ...  \n",
       "3  like strong smooth flavor coffe tri other keep...  \n",
       "4  mae ploy sweet chili sauc becom standard condi...  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_10k.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_10k.drop('index',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ProductId</th>\n",
       "      <th>UserId</th>\n",
       "      <th>ProfileName</th>\n",
       "      <th>HelpfulnessNumerator</th>\n",
       "      <th>HelpfulnessDenominator</th>\n",
       "      <th>Time</th>\n",
       "      <th>Summary</th>\n",
       "      <th>Text</th>\n",
       "      <th>CleanedText</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>B00002Z754</td>\n",
       "      <td>A29Z5PI9BW2PU3</td>\n",
       "      <td>Robbie</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>961718400</td>\n",
       "      <td>Great Product</td>\n",
       "      <td>This was a really good idea and the final prod...</td>\n",
       "      <td>realli good idea final product outstand use de...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>B00002Z754</td>\n",
       "      <td>A3B8RCEI0FXFI6</td>\n",
       "      <td>B G Chase</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>962236800</td>\n",
       "      <td>WOW Make your own 'slickers' !</td>\n",
       "      <td>I just received my shipment and could hardly w...</td>\n",
       "      <td>receiv shipment could hard wait tri product lo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>B0000EIE2Z</td>\n",
       "      <td>A3M174IC0VXOS2</td>\n",
       "      <td>Gail Cooke</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>1075420800</td>\n",
       "      <td>BEST BLUEBERRIES</td>\n",
       "      <td>In the winter when fresh blueberries exceed ou...</td>\n",
       "      <td>winter fresh blueberri exceed food budget dri ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B0000VYKXC</td>\n",
       "      <td>A3L88AH6MABEDL</td>\n",
       "      <td>Javier Denison</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1081209600</td>\n",
       "      <td>Best coffee I've tasted. I try others but alwa...</td>\n",
       "      <td>If you like a strong but smooth flavorful coff...</td>\n",
       "      <td>like strong smooth flavor coffe tri other keep...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>B00016UX0K</td>\n",
       "      <td>AF1PV3DIC0XM7</td>\n",
       "      <td>Robert Ashton</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1081555200</td>\n",
       "      <td>Classic Condiment</td>\n",
       "      <td>Mae Ploy Sweet Chili Sauce is becoming a stand...</td>\n",
       "      <td>mae ploy sweet chili sauc becom standard condi...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    ProductId          UserId     ProfileName  HelpfulnessNumerator  \\\n",
       "0  B00002Z754  A29Z5PI9BW2PU3          Robbie                     7   \n",
       "1  B00002Z754  A3B8RCEI0FXFI6       B G Chase                    10   \n",
       "2  B0000EIE2Z  A3M174IC0VXOS2      Gail Cooke                     3   \n",
       "3  B0000VYKXC  A3L88AH6MABEDL  Javier Denison                     0   \n",
       "4  B00016UX0K   AF1PV3DIC0XM7   Robert Ashton                     1   \n",
       "\n",
       "   HelpfulnessDenominator        Time  \\\n",
       "0                       7   961718400   \n",
       "1                      10   962236800   \n",
       "2                       3  1075420800   \n",
       "3                       0  1081209600   \n",
       "4                       2  1081555200   \n",
       "\n",
       "                                             Summary  \\\n",
       "0                                      Great Product   \n",
       "1                     WOW Make your own 'slickers' !   \n",
       "2                                   BEST BLUEBERRIES   \n",
       "3  Best coffee I've tasted. I try others but alwa...   \n",
       "4                                  Classic Condiment   \n",
       "\n",
       "                                                Text  \\\n",
       "0  This was a really good idea and the final prod...   \n",
       "1  I just received my shipment and could hardly w...   \n",
       "2  In the winter when fresh blueberries exceed ou...   \n",
       "3  If you like a strong but smooth flavorful coff...   \n",
       "4  Mae Ploy Sweet Chili Sauce is becoming a stand...   \n",
       "\n",
       "                                         CleanedText  \n",
       "0  realli good idea final product outstand use de...  \n",
       "1  receiv shipment could hard wait tri product lo...  \n",
       "2  winter fresh blueberri exceed food budget dri ...  \n",
       "3  like strong smooth flavor coffe tri other keep...  \n",
       "4  mae ploy sweet chili sauc becom standard condi...  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_10k.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_1, X_test, y_1, y_test = train_test_split(filtered_10k, label, test_size=0.2, random_state=50)\n",
    "X_tr, X_cv, y_tr, y_cv = train_test_split(X_1,y_1, test_size=0.25, random_state=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8000, 9) 8000\n",
      "(6000, 9) 6000\n"
     ]
    }
   ],
   "source": [
    "print(X_1.shape,len(y_1))\n",
    "print(X_tr.shape,len(y_tr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Bag Of Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "count_vector=CountVectorizer(ngram_range=(1,2))\n",
    "vocab=count_vector.fit(X_tr.CleanedText)\n",
    "\n",
    "bow_train=count_vector.transform(X_tr.CleanedText)\n",
    "bow_cv=count_vector.transform(X_cv.CleanedText)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6000, 139286)\n",
      "(2000, 139286)\n"
     ]
    }
   ],
   "source": [
    "print(bow_train.get_shape())\n",
    "print(bow_cv.get_shape())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "bow_train_array=bow_train.toarray()\n",
    "bow_cv_array=bow_cv.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=1, n_neighbors=1, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model=KNeighborsClassifier(n_neighbors=1)\n",
    "model.fit(bow_train_array,y_tr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "prediction=model.predict(bow_cv_array)\n",
    "cnf_matrix=confusion_matrix(y_cv, prediction, labels=None, sample_weight=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    " \n",
    "def fit_and_cv(X_train,y_train,X_test,y_test):\n",
    "    nbrs_list=list(range(1,11)) \n",
    "    for i in nbrs_list:\n",
    "        model=KNeighborsClassifier(n_neighbors=i)\n",
    "        model.fit(X_train,y_train)\n",
    "        print(\"For n_neighbors =\",str(i))\n",
    "        prediction=model.predict(X_test)\n",
    "        accuracy = accuracy_score(y_test, prediction, normalize=True) * float(100)\n",
    "        print (\"\\nmodel accuracy is {} %\".format(accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For n_neighbors ="
     ]
    }
   ],
   "source": [
    "fit_and_cv(bow_train_array,y_tr,bow_cv_array,y_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4000, 53711)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_data_count.shape\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simple CV for practice:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_1, X_test, y_1, y_test = train_test_split(filtered_data_count, filtered_4k.Sentiment, test_size=0.3, random_state=50)\n",
    "X_tr, X_cv, y_tr, y_cv = train_test_split(X_1,y_1, test_size=0.3, random_state=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For n_neighbors = 1\n",
      "\n",
      "model accuracy is 90.23809523809524 %\n",
      "For n_neighbors = 2\n",
      "\n",
      "model accuracy is 73.33333333333333 %\n",
      "For n_neighbors = 3\n",
      "\n",
      "model accuracy is 81.9047619047619 %\n",
      "For n_neighbors = 4\n",
      "\n",
      "model accuracy is 70.23809523809523 %\n",
      "For n_neighbors = 5\n",
      "\n",
      "model accuracy is 87.14285714285714 %\n",
      "For n_neighbors = 6\n",
      "\n",
      "model accuracy is 86.19047619047619 %\n",
      "For n_neighbors = 7\n",
      "\n",
      "model accuracy is 84.04761904761905 %\n",
      "For n_neighbors = 8\n",
      "\n",
      "model accuracy is 84.04761904761905 %\n",
      "For n_neighbors = 9\n",
      "\n",
      "model accuracy is 84.04761904761905 %\n",
      "For n_neighbors = 10\n",
      "\n",
      "model accuracy is 83.92857142857143 %\n"
     ]
    }
   ],
   "source": [
    "nbrs_list=list(range(1,11))\n",
    "for i in nbrs_list:\n",
    "    print(\"For n_neighbors =\",str(i))\n",
    "    fit_model(X_tr,y_tr,X_cv,y_cv,i)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "def model_with_cv(X_train,y_train,X_test,y_test,cv):\n",
    "    nbrs_list=list(range(1,11,2))\n",
    "\n",
    "# empty list that will hold cv scores\n",
    "    cv_scores = []\n",
    "\n",
    "# perform 10-fold cross validation\n",
    "    for k in nbrs_list:\n",
    "        knn = KNeighborsClassifier(n_neighbors=k)\n",
    "        scores = cross_validation.cross_val_score(knn, X_train, y_train, cv=cv, scoring='accuracy')\n",
    "        cv_scores.append(scores.mean())\n",
    "\n",
    "# changing to misclassification error\n",
    "    MSE = [1 - x for x in cv_scores]\n",
    "\n",
    "# determining best k\n",
    "    optimal_k = nbrs_list[MSE.index(min(MSE))]\n",
    "    print('\\nThe optimal number of neighbors is %d.' % optimal_k)\n",
    "\n",
    "# plot misclassification error vs k \n",
    "    plt.plot(nbrs_list, MSE)\n",
    "\n",
    "    for xy in zip(nbrs_list, np.round(MSE,3)):\n",
    "        plt.annotate('(%s, %s)' % xy, xy=xy, textcoords='data')\n",
    "\n",
    "    plt.xlabel('Number of Neighbors K')\n",
    "    plt.ylabel('Misclassification Error')\n",
    "    plt.show()\n",
    "\n",
    "    print(\"the misclassification error for each k value is : \", np.round(MSE,3))\n",
    "\n",
    "    knn_optimal = KNeighborsClassifier(n_neighbors=optimal_k)\n",
    "\n",
    "# fitting the model\n",
    "    knn_optimal.fit(X_train, y_train)\n",
    "\n",
    "# predict the response\n",
    "    pred = knn_optimal.predict(X_test)\n",
    "\n",
    "# evaluate accuracy\n",
    "    acc = accuracy_score(y_test, pred) * 100\n",
    "    print('\\nThe accuracy of the knn classifier for k = %d is %f%%' % (optimal_k, acc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The optimal number of neighbors is 1.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaoAAAEKCAYAAABDkxEYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nyhhjjFezRGWMMcarWaIyxhjj1f4fLSHK6bzKJfIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2d2aa9d31d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the misclassification error for each k value is :  [ 0.072  0.164  0.174  0.19   0.153]\n",
      "\n",
      "The accuracy of the knn classifier for k = 1 is 95.666667%\n"
     ]
    }
   ],
   "source": [
    "model_with_cv(X_1,y_1,X_test,y_test,10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TF-IDF intialization and dimension creation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the type of count vectorizer  <class 'scipy.sparse.csr.csr_matrix'>\n",
      "the shape of out TF-IDF vectorizer  (4000, 53711)\n"
     ]
    }
   ],
   "source": [
    "tf_idf_vect = TfidfVectorizer(ngram_range=(1,2))\n",
    "filtered_tf_idf = tf_idf_vect.fit_transform(filtered_4k.CleanedText)\n",
    "print(\"the type of count vectorizer \",type(filtered_tf_idf))\n",
    "print(\"the shape of out TF-IDF vectorizer \",filtered_tf_idf.get_shape())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_tf_idf=filtered_tf_idf.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_1, X_test, y_1, y_test = train_test_split(filtered_tf_idf, filtered_4k.Sentiment, test_size=0.3, random_state=50)\n",
    "#X_tr, X_cv, y_tr, y_cv = train_test_split(X_1,y_1, test_size=0.3, random_state=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The optimal number of neighbors is 1.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaMAAAEKCAYAAAC/hjrSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2d2aaf291d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the misclassification error for each k value is :  [ 0.066  0.149  0.129  0.136  0.13 ]\n",
      "\n",
      "The accuracy of the knn classifier for k = 1 is 95.000000%\n"
     ]
    }
   ],
   "source": [
    "model_with_cv(X_1,y_1,X_test,y_test,10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Wieghted Word2Vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "list_of_sent=[]\n",
    "for sent in filtered_4k.CleanedText:\n",
    "    list_of_sent.append(sent.split())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4000\n",
      "['realli', 'good', 'idea', 'final', 'product', 'outstand', 'use', 'decal', 'car', 'window', 'everybodi', 'ask', 'bought', 'decal', 'made', 'two', 'thumb']\n"
     ]
    }
   ],
   "source": [
    "print(len(list_of_sent))\n",
    "print(list_of_sent[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4000\n",
      "50\n"
     ]
    }
   ],
   "source": [
    "w2v_model=Word2Vec(list_of_sent,min_count=5,size=50, workers=4)\n",
    "w2v_words = list(w2v_model.wv.vocab)\n",
    "# average Word2Vec\n",
    "# compute average word2vec for each review.\n",
    "sent_vectors = []; # the avg-w2v for each sentence/review is stored in this list\n",
    "for sent in list_of_sent: # for each review/sentence\n",
    "    sent_vec = np.zeros(50) # as word vectors are of zero length\n",
    "    cnt_words =0; # num of words with a valid vector in the sentence/review\n",
    "    for word in sent: # for each word in a review/sentence\n",
    "        if word in w2v_words:\n",
    "            vec = w2v_model.wv[word]\n",
    "            sent_vec += vec\n",
    "            cnt_words += 1\n",
    "    if cnt_words != 0:\n",
    "        sent_vec /= cnt_words\n",
    "    sent_vectors.append(sent_vec)\n",
    "print(len(sent_vectors))\n",
    "print(len(sent_vectors[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "list"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(sent_vectors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_1, X_test, y_1, y_test = train_test_split(sent_vectors, filtered_4k.Sentiment, test_size=0.3, random_state=50)\n",
    "#X_tr, X_cv, y_tr, y_cv = train_test_split(X_1,y_1, test_size=0.3, random_state=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The optimal number of neighbors is 1.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaoAAAEKCAYAAABDkxEYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBDVLVMYYY4La/wf93EsPBr2WfAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2d1fc91ddd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the misclassification error for each k value is :  [ 0.099  0.193  0.163  0.168  0.164]\n",
      "\n",
      "The accuracy of the knn classifier for k = 1 is 92.500000%\n"
     ]
    }
   ],
   "source": [
    "model_with_cv(X_1,y_1,X_test,y_test,10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Wieghted TF-IDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tfidf_feat = tf_idf_vect.get_feature_names() # tfidf words/col-names\n",
    "# final_tf_idf is the sparse matrix with row= sentence, col=word and cell_val = tfidf\n",
    "\n",
    "tfidf_sent_vectors = []; # the tfidf-w2v for each sentence/review is stored in this list\n",
    "row=0;\n",
    "for sent in list_of_sent: # for each review/sentence \n",
    "    sent_vec = np.zeros(50) # as word vectors are of zero length\n",
    "    weight_sum =0; # num of words with a valid vector in the sentence/review\n",
    "    for word in sent: # for each word in a review/sentence\n",
    "        if word in w2v_words:\n",
    "            vec = w2v_model.wv[word]\n",
    "            # obtain the tf_idfidf of a word in a sentence/review\n",
    "            tf_idf = filtered_tf_idf[row, tfidf_feat.index(word)]\n",
    "            sent_vec += (vec * tf_idf)\n",
    "            weight_sum += tf_idf\n",
    "    if weight_sum != 0:\n",
    "        sent_vec /= weight_sum\n",
    "    tfidf_sent_vectors.append(sent_vec)\n",
    "    row += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "list"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(tfidf_sent_vectors)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_1, X_test, y_1, y_test = train_test_split(tfidf_sent_vectors, filtered_4k.Sentiment, test_size=0.3, random_state=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The optimal number of neighbors is 1.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaoAAAEKCAYAAABDkxEYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2d19522e080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the misclassification error for each k value is :  [ 0.098  0.193  0.154  0.166  0.155]\n",
      "\n",
      "The accuracy of the knn classifier for k = 1 is 92.833333%\n"
     ]
    }
   ],
   "source": [
    "model_with_cv(X_1,y_1,X_test,y_test,10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusion : The best K is 1 for all 4 methods , but accuracy of BOW (95.66) is slightly higher than others. Another observation is \"wieghted word2vec\" and \"wieghted TF-IDF took significantly less time than other two."
   ]
  },
  {
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}