Sentimental_analysis_Airlines+Tweets.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to\n",
      "[nltk_data]     C:\\Users\\pothulas\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n",
      "[nltk_data] Downloading package perluniprops to\n",
      "[nltk_data]     C:\\Users\\pothulas\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package perluniprops is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import time\n",
    "import os\n",
    "try:\n",
    "    import nltk\n",
    "except:\n",
    "    !pip install nltk\n",
    "from nltk import word_tokenize,sent_tokenize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "os.chdir(r'\\\\SRVA\\Homes$\\pothulas\\Desktop\\DataMining\\Project')\n",
    "df =pd.read_csv('Tweets.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>tweet_id</th>\n",
       "      <th>airline_sentiment</th>\n",
       "      <th>airline_sentiment_confidence</th>\n",
       "      <th>negativereason</th>\n",
       "      <th>negativereason_confidence</th>\n",
       "      <th>airline</th>\n",
       "      <th>airline_sentiment_gold</th>\n",
       "      <th>name</th>\n",
       "      <th>negativereason_gold</th>\n",
       "      <th>retweet_count</th>\n",
       "      <th>text</th>\n",
       "      <th>tweet_coord</th>\n",
       "      <th>tweet_created</th>\n",
       "      <th>tweet_location</th>\n",
       "      <th>user_timezone</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>570306133677760513</td>\n",
       "      <td>neutral</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>cairdin</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:35:52 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Eastern Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>570301130888122368</td>\n",
       "      <td>positive</td>\n",
       "      <td>0.3486</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>jnardino</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:15:59 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>570301083672813571</td>\n",
       "      <td>neutral</td>\n",
       "      <td>0.6837</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>yvonnalynn</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:15:48 -0800</td>\n",
       "      <td>Lets Play</td>\n",
       "      <td>Central Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>570301031407624196</td>\n",
       "      <td>negative</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>Bad Flight</td>\n",
       "      <td>0.7033</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>jnardino</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:15:36 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>570300817074462722</td>\n",
       "      <td>negative</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>Can't Tell</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>jnardino</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:14:45 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             tweet_id airline_sentiment  airline_sentiment_confidence  \\\n",
       "0  570306133677760513           neutral                        1.0000   \n",
       "1  570301130888122368          positive                        0.3486   \n",
       "2  570301083672813571           neutral                        0.6837   \n",
       "3  570301031407624196          negative                        1.0000   \n",
       "4  570300817074462722          negative                        1.0000   \n",
       "\n",
       "  negativereason  negativereason_confidence         airline  \\\n",
       "0            NaN                        NaN  Virgin America   \n",
       "1            NaN                     0.0000  Virgin America   \n",
       "2            NaN                        NaN  Virgin America   \n",
       "3     Bad Flight                     0.7033  Virgin America   \n",
       "4     Can't Tell                     1.0000  Virgin America   \n",
       "\n",
       "  airline_sentiment_gold        name negativereason_gold  retweet_count  \\\n",
       "0                    NaN     cairdin                 NaN              0   \n",
       "1                    NaN    jnardino                 NaN              0   \n",
       "2                    NaN  yvonnalynn                 NaN              0   \n",
       "3                    NaN    jnardino                 NaN              0   \n",
       "4                    NaN    jnardino                 NaN              0   \n",
       "\n",
       "                                                text tweet_coord  \\\n",
       "0                @VirginAmerica What @dhepburn said.         NaN   \n",
       "1  @VirginAmerica plus you've added commercials t...         NaN   \n",
       "2  @VirginAmerica I didn't today... Must mean I n...         NaN   \n",
       "3  @VirginAmerica it's really aggressive to blast...         NaN   \n",
       "4  @VirginAmerica and it's a really big bad thing...         NaN   \n",
       "\n",
       "               tweet_created tweet_location               user_timezone  \n",
       "0  2015-02-24 11:35:52 -0800            NaN  Eastern Time (US & Canada)  \n",
       "1  2015-02-24 11:15:59 -0800            NaN  Pacific Time (US & Canada)  \n",
       "2  2015-02-24 11:15:48 -0800      Lets Play  Central Time (US & Canada)  \n",
       "3  2015-02-24 11:15:36 -0800            NaN  Pacific Time (US & Canada)  \n",
       "4  2015-02-24 11:14:45 -0800            NaN  Pacific Time (US & Canada)  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 14640 entries, 0 to 14639\n",
      "Data columns (total 15 columns):\n",
      "tweet_id                        14640 non-null int64\n",
      "airline_sentiment               14640 non-null object\n",
      "airline_sentiment_confidence    14640 non-null float64\n",
      "negativereason                  9178 non-null object\n",
      "negativereason_confidence       10522 non-null float64\n",
      "airline                         14640 non-null object\n",
      "airline_sentiment_gold          40 non-null object\n",
      "name                            14640 non-null object\n",
      "negativereason_gold             32 non-null object\n",
      "retweet_count                   14640 non-null int64\n",
      "text                            14640 non-null object\n",
      "tweet_coord                     1019 non-null object\n",
      "tweet_created                   14640 non-null object\n",
      "tweet_location                  9907 non-null object\n",
      "user_timezone                   9820 non-null object\n",
      "dtypes: float64(2), int64(2), object(11)\n",
      "memory usage: 1.7+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14640, 15)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cols = ['tweet_id','negativereason_confidence','tweet_created','airline_sentiment_confidence','negativereason','airline_sentiment_gold','name','negativereason_gold','retweet_count','tweet_coord']\n",
    "df1 = df.drop(cols, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>airline_sentiment</th>\n",
       "      <th>airline</th>\n",
       "      <th>text</th>\n",
       "      <th>tweet_location</th>\n",
       "      <th>user_timezone</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13336</th>\n",
       "      <td>negative</td>\n",
       "      <td>American</td>\n",
       "      <td>@AmericanAir flight US1562 from RIC2DFW was Ca...</td>\n",
       "      <td>Richmond, VA</td>\n",
       "      <td>Central Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8229</th>\n",
       "      <td>negative</td>\n",
       "      <td>Delta</td>\n",
       "      <td>@JetBlue it is absurd that I have waited more ...</td>\n",
       "      <td>Miami | Puerto Rico</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8459</th>\n",
       "      <td>positive</td>\n",
       "      <td>Delta</td>\n",
       "      <td>@JetBlue btw, her name was Samantha and she wo...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Eastern Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3944</th>\n",
       "      <td>negative</td>\n",
       "      <td>United</td>\n",
       "      <td>@united are my bags here yet?  They were at Pa...</td>\n",
       "      <td>All Over The World</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7359</th>\n",
       "      <td>negative</td>\n",
       "      <td>Delta</td>\n",
       "      <td>@JetBlue how about free wifi on flt 1254 out o...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Atlantic Time (Canada)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      airline_sentiment   airline  \\\n",
       "13336          negative  American   \n",
       "8229           negative     Delta   \n",
       "8459           positive     Delta   \n",
       "3944           negative    United   \n",
       "7359           negative     Delta   \n",
       "\n",
       "                                                    text  \\\n",
       "13336  @AmericanAir flight US1562 from RIC2DFW was Ca...   \n",
       "8229   @JetBlue it is absurd that I have waited more ...   \n",
       "8459   @JetBlue btw, her name was Samantha and she wo...   \n",
       "3944   @united are my bags here yet?  They were at Pa...   \n",
       "7359   @JetBlue how about free wifi on flt 1254 out o...   \n",
       "\n",
       "             tweet_location               user_timezone  \n",
       "13336          Richmond, VA  Central Time (US & Canada)  \n",
       "8229   Miami | Puerto Rico                          NaN  \n",
       "8459                    NaN  Eastern Time (US & Canada)  \n",
       "3944     All Over The World                         NaN  \n",
       "7359                    NaN      Atlantic Time (Canada)  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "United            3822\n",
       "US Airways        2913\n",
       "American          2759\n",
       "Southwest         2420\n",
       "Delta             2222\n",
       "Virgin America     504\n",
       "Name: airline, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1['airline'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "negative    9178\n",
       "neutral     3099\n",
       "positive    2363\n",
       "Name: airline_sentiment, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1['airline_sentiment'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"@VirginAmerica it was amazing, and arrived an hour early. You're too good to me.\""
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['text'][9]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import re\n",
    "import string\n",
    "def cleaning(tweet):\n",
    "  cleanr = re.compile('<.*?>')      # logic: Removing HTML tags by using regular expressions\n",
    "  clean1 = re.sub(cleanr, '', tweet)\n",
    "  clean2 = re.sub(r'(@[A-Za-z0-9]+)|([^0-9A-Za-z \\t])|(\\w+:\\/\\/\\S+)', \"\", clean1)  #Removing Url('.' with continous string on eitherside if present is removed )\n",
    "  clean3 = str(clean2).replace('\\xa0',' ')\n",
    "  clean4 = str(clean3).lower()              # converting all tweets to lower case to create uniformity for model fitting of words\n",
    "  translation = str.maketrans('', '', string.punctuation)\n",
    "  clean5 = clean4.translate(translation)        # Removing all special characters\n",
    "  clean6 = clean5.replace('…', ' ')\n",
    "  return clean6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "stop=pd.read_csv(\"stop_words.txt\", header=None)\n",
    "stop = set(stop[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df1['text'] = df1.text.apply(lambda x: cleaning(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from nltk.tokenize import word_tokenize\n",
    "df1['text'] = df1.text.apply(lambda x: word_tokenize(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df1.text = df1['text'].apply(lambda x: [word for word in x if word not in stop])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'@SouthwestAir #DestinationDragons @Imaginedragons Scavenger Hunt Vegas, BE READY!!! http://t.co/vHgkiTzSaw'"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['text'][4785]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'@united Is there a direct number to call to request a refund?  Its telling me its not valid as I changed the ticket.  No number on site'"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['text'][4255]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['direct',\n",
       " 'number',\n",
       " 'call',\n",
       " 'request',\n",
       " 'refund',\n",
       " 'telling',\n",
       " 'valid',\n",
       " 'changed',\n",
       " 'ticket',\n",
       " 'number',\n",
       " 'site']"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1['text'][4255]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfgAAAFZCAYAAAB0aD00AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGrjQmmsgAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28a96e0ceb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x ='airline', data = df1)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfgAAAFXCAYAAABOYlxEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28a91966c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x ='airline_sentiment', data = df1)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "negative    9178\n",
       "neutral     3099\n",
       "positive    2363\n",
       "Name: airline_sentiment, dtype: int64"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['airline_sentiment'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcwAAAFLCAYAAABbbrzmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIi6gIVJRETUBSxMIiKiLvh/lwEL9hRNrhsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28a96dbe748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = df['airline_sentiment'].value_counts()\n",
    "x.tolist()\n",
    "Tweet_Sentiment = ['Negative', 'Neutral','Positive']\n",
    "piechart = plt.pie(x,  labels = Tweet_Sentiment, autopct = '%1.1f%%')\n",
    "plt.title('Tweets Sentiment Percentage', fontsize = 15, fontweight = 'bold')\n",
    "plt.axis('equal')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsYAAAGBCAYAAABl3Xf8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkiTAYCxJkiQBBmNJkiQJMBhLkiRJAPw/n+bn1ldsJaAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28a96dba8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "a4_dims = (11.7, 6.27)\n",
    "fig, ax = plt.subplots(figsize=a4_dims)\n",
    "sns.countplot(x='airline', hue=\"airline_sentiment\", data=df1)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from nltk.tokenize.moses import MosesDetokenizer          # Detokenizing the tweets back to string using Mosesdetokenizer\n",
    "detokenizer = MosesDetokenizer()\n",
    "df1.text = df1.text.apply(lambda x: detokenizer.detokenize(x, return_str=True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>airline_sentiment</th>\n",
       "      <th>airline</th>\n",
       "      <th>text</th>\n",
       "      <th>tweet_location</th>\n",
       "      <th>user_timezone</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>neutral</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td></td>\n",
       "      <td>NaN</td>\n",
       "      <td>Eastern Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>youve commercials experience tacky</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>neutral</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>didnt today trip</td>\n",
       "      <td>Lets Play</td>\n",
       "      <td>Central Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>aggressive blast obnoxious entertainment guest...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>big bad thing</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment         airline  \\\n",
       "0           neutral  Virgin America   \n",
       "1          positive  Virgin America   \n",
       "2           neutral  Virgin America   \n",
       "3          negative  Virgin America   \n",
       "4          negative  Virgin America   \n",
       "\n",
       "                                                text tweet_location  \\\n",
       "0                                                               NaN   \n",
       "1                 youve commercials experience tacky            NaN   \n",
       "2                                   didnt today trip      Lets Play   \n",
       "3  aggressive blast obnoxious entertainment guest...            NaN   \n",
       "4                                      big bad thing            NaN   \n",
       "\n",
       "                user_timezone  \n",
       "0  Eastern Time (US & Canada)  \n",
       "1  Pacific Time (US & Canada)  \n",
       "2  Central Time (US & Canada)  \n",
       "3  Pacific Time (US & Canada)  \n",
       "4  Pacific Time (US & Canada)  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "string = []\n",
    "l = df1.text.apply(lambda x: string.extend([x]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "all_tweets_string = ' '.join(str(e) for e in string)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "f = open('all_tweets_string.txt','w')\n",
    "f.write(all_tweets_string)\n",
    "f.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "United            3822\n",
       "US Airways        2913\n",
       "American          2759\n",
       "Southwest         2420\n",
       "Delta             2222\n",
       "Virgin America     504\n",
       "Name: airline, dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['airline'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### For Plotting Word clouds - Data of different Airlines seperated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "string = []\n",
    "l = df_United.text.apply(lambda x: string.extend([x]))\n",
    "United_tweets_string = ' '.join(str(e) for e in string)\n",
    "f = open('United_tweets_string.txt','w')\n",
    "f.write(United_tweets_string)\n",
    "f.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "string = []\n",
    "l = df_US_Airways.text.apply(lambda x: string.extend([x]))\n",
    "USairways_tweets_string = ' '.join(str(e) for e in string)\n",
    "f = open('USairways_tweets_string.txt','w')\n",
    "f.write(USairways_tweets_string)\n",
    "f.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "#### Dividing into Testing & Training sets:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(df1['text'], df1['airline_sentiment'], test_size =0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = pd.DataFrame(X_train)\n",
    "X_test = pd.DataFrame(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "negative    2823\n",
       "neutral      875\n",
       "positive     694\n",
       "Name: airline_sentiment, dtype: int64"
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model Implementation:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.cross_validation import cross_val_score      # Importing Cross Validation from sklearn\n",
    "count_vect = CountVectorizer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = X_train['text'].astype(str).values.tolist()\n",
    "X_test = X_test['text'].astype(str).values.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "z = count_vect.fit_transform(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "w = count_vect.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression      # Importing Logistic Regression Function\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the logistic Regression classifier with 10 fold cross-validation is 0.695500045828\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.17950701316197712"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "start = time.time()              # Time function used to evaluate the execution run time for training logistic model\n",
    "\n",
    "logreg =  LogisticRegression(class_weight='balanced', multi_class='multinomial', solver='lbfgs',C = 0.5125)\n",
    "\n",
    "logreg_scores = cross_val_score(logreg, X1,y, cv=10, scoring='accuracy')\n",
    "print('The accuracy of the logistic Regression classifier with 10 fold cross-validation is ' + str(logreg_scores.mean()))\n",
    "\n",
    "end = time.time()\n",
    "logreg_time = (end - start)/60    # Time converted to Minutes\n",
    "logreg_time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best set of parameters found:\n",
      "{'C': 0.51250000000000007}\n",
      "\n",
      "Scores on training set:\n",
      "0.683 (+/-0.072) for {'C': 0.10000000000000001}\n",
      "0.695 (+/-0.068) for {'C': 0.51250000000000007}\n",
      "0.693 (+/-0.066) for {'C': 0.92500000000000004}\n",
      "0.693 (+/-0.065) for {'C': 1.3375000000000001}\n",
      "0.693 (+/-0.064) for {'C': 1.7500000000000002}\n",
      "0.692 (+/-0.065) for {'C': 2.1625000000000001}\n",
      "0.692 (+/-0.066) for {'C': 2.5750000000000002}\n",
      "0.691 (+/-0.064) for {'C': 2.9875000000000003}\n",
      "0.690 (+/-0.065) for {'C': 3.4000000000000004}\n",
      "0.690 (+/-0.066) for {'C': 3.8125000000000004}\n",
      "0.689 (+/-0.065) for {'C': 4.2249999999999996}\n",
      "0.689 (+/-0.065) for {'C': 4.6375000000000002}\n",
      "0.688 (+/-0.064) for {'C': 5.0499999999999998}\n",
      "0.688 (+/-0.064) for {'C': 5.4625000000000004}\n",
      "0.687 (+/-0.064) for {'C': 5.875}\n",
      "0.687 (+/-0.065) for {'C': 6.2875000000000005}\n",
      "0.686 (+/-0.062) for {'C': 6.7000000000000002}\n",
      "0.686 (+/-0.064) for {'C': 7.1124999999999998}\n",
      "0.685 (+/-0.062) for {'C': 7.5250000000000004}\n",
      "0.685 (+/-0.063) for {'C': 7.9375}\n",
      "0.685 (+/-0.064) for {'C': 8.3499999999999996}\n",
      "0.684 (+/-0.063) for {'C': 8.7625000000000011}\n",
      "0.685 (+/-0.066) for {'C': 9.1750000000000007}\n",
      "0.684 (+/-0.064) for {'C': 9.5875000000000004}\n",
      "0.684 (+/-0.062) for {'C': 10.0}\n",
      "\n",
      "The hyperparameter tuning took:  266.56325221061707 seconds\n"
     ]
    }
   ],
   "source": [
    "start_time = time.time()\n",
    "\n",
    "logreg = LogisticRegression(class_weight='balanced', multi_class='multinomial', solver='lbfgs',random_state=0)\n",
    "param_grid = {'C':np.linspace(0.1,10, 25)}\n",
    "\n",
    "grid = GridSearchCV(logreg, param_grid, cv=10, scoring='accuracy')\n",
    "grid.fit(X1, y)\n",
    "\n",
    "print(\"Best set of parameters found:\")\n",
    "print(grid.best_params_)\n",
    "print()\n",
    "\n",
    "print(\"Scores on training set:\")\n",
    "means = grid.cv_results_['mean_test_score']\n",
    "stds = grid.cv_results_['std_test_score']\n",
    "for mean, std, params in zip(means, stds, grid.cv_results_['params']):\n",
    "    print(\"%0.3f (+/-%0.03f) for %r\"\n",
    "          % (mean, std * 2, params))\n",
    "print()\n",
    "print(\"The hyperparameter tuning took:  %s seconds\" % (time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "70.651183970856096"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logreg.fit(z, y_train.values.ravel())\n",
    "logreg_accuracy = logreg.score(w, y_test)*100    # Accuracy of the Logistic Regression Model on the test set\n",
    "logreg_accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "59.638574393140409"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.cross_validation import cross_val_score      # Importing Cross Validation from sklearn\n",
    "\n",
    "X1 =df1['text'].astype(str).values.tolist()\n",
    "X1 = count_vect.transform(X1)\n",
    "y = df1['airline_sentiment']\n",
    "\n",
    "Log_reg_cross_val = (cross_val_score(logreg, X, y, cv=10, scoring='accuracy').mean())\n",
    "Log_reg_cross_val*100                                      # Finding the mean accuracy of 10 cases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59.6385743931\n"
     ]
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer   # Importing TfidfVectorizer\n",
    "Tfid_vect = TfidfVectorizer(min_df=1,max_features = 100)\n",
    "\n",
    "X2 =df1['text'].astype(str).values.tolist()            # Fitting or transforming text data to Tfidf Vectorizer\n",
    "X2 = Tfid_vect.fit_transform(X2)\n",
    "y = df1['airline_sentiment']\n",
    "\n",
    "Log_reg_Tfid_cross_val = (cross_val_score(logreg, X, y, cv=10, scoring='accuracy').mean())\n",
    "print(Log_reg_Tfid_cross_val*100)                                                     "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True: ['negative' 'negative' 'neutral' 'negative' 'negative' 'positive'\n",
      " 'negative' 'negative' 'neutral' 'negative' 'positive' 'negative'\n",
      " 'negative' 'negative' 'positive' 'negative' 'negative' 'neutral'\n",
      " 'positive' 'negative' 'neutral' 'negative' 'neutral' 'negative' 'positive']\n"
     ]
    }
   ],
   "source": [
    "## Printing first 25 original and predicted classes for the tweets\n",
    "print('True:', y_test.values[0:25])\n",
    "#print('False:', y_pred_class[0:25])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "forest = RandomForestClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the Random forest classifier with 10 fold cross-validation is 0.684838377148\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7135929544766744"
      ]
     },
     "execution_count": 198,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "start_forest = time.time()\n",
    "\n",
    "forest_scores = cross_val_score(forest, X1, y, cv=10, scoring='accuracy')\n",
    "print('The accuracy of the Random forest classifier with 10 fold cross-validation is ' + str(forest_scores.mean()))\n",
    "\n",
    "end_forest = time.time()\n",
    "forest_time = (end_forest-start_forest)/60\n",
    "forest_time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.96945745511319281"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest.fit(z, y_train.values.ravel())\n",
    "forest.score(z,y_train)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "70.104735883424411"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest_accuracy = forest.score(w, y_test)*100           ## Testing Accuracy for Random Forest Classifier\n",
    "forest_accuracy "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the SVM classifier with 10 fold cross-validation is 0.71740541262\n"
     ]
    }
   ],
   "source": [
    "svc_2=SVC(kernel='rbf',C = 10, gamma = 0.05,)\n",
    "svc2_scores = cross_val_score(svc,z, y_train.values.ravel(), cv=10, scoring='accuracy')\n",
    "print('The accuracy of the SVM classifier with 10 fold cross-validation is ' + str(svc2_scores.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the SVM classifier with 10 fold cross-validation is 0.646380536186\n"
     ]
    }
   ],
   "source": [
    "svc_2=SVC(kernel='rbf')\n",
    "svc2_scores = cross_val_score(svc,X2,y, cv=10, scoring='accuracy')\n",
    "print('The accuracy of the SVM classifier with 10 fold cross-validation is ' + str(svc2_scores.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['negative', 'negative', 'positive', ..., 'negative', 'negative',\n",
       "       'positive'], dtype=object)"
      ]
     },
     "execution_count": 181,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc_2=SVC(kernel='rbf',C = 10, gamma = 0.05,)\n",
    "svc_2.fit(z, y_train.values.ravel())\n",
    "svc_2.predict(w)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(C=10, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape=None, degree=3, gamma=0.05, kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 207,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc_2.fit(z, y_train.values.ravel())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predicted = svc_2.predict(w)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.73952641165755917"
      ]
     },
     "execution_count": 209,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc_2.score(w,y_test)   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Anaconda\\lib\\site-packages\\sklearn\\svm\\base.py:552: ChangedBehaviorWarning: The decision_function_shape default value will change from 'ovo' to 'ovr' in 0.19. This will change the shape of the decision function returned by SVC.\n",
      "  \"SVC.\", ChangedBehaviorWarning)\n"
     ]
    }
   ],
   "source": [
    "y_score = svc_2.fit(z, y_train.values.ravel()).decision_function(w)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The system cannot find the path specified.\n"
     ]
    }
   ],
   "source": [
    "!pip install mlxtend"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cm_data = confusion_matrix(y_test, predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[2394,  316,  113],\n",
       "       [ 272,  511,   92],\n",
       "       [ 160,  191,  343]])"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cm_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14640, 15)"
      ]
     },
     "execution_count": 194,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best set of parameters found:\n",
      "{'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 500, 'n_estimators': 10}\n",
      "\n",
      "Scores on training set:\n",
      "0.591 (+/-0.044) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 200, 'n_estimators': 10}\n",
      "0.586 (+/-0.039) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 200, 'n_estimators': 50}\n",
      "0.587 (+/-0.039) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 200, 'n_estimators': 100}\n",
      "0.593 (+/-0.041) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 500, 'n_estimators': 10}\n",
      "0.588 (+/-0.045) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 500, 'n_estimators': 50}\n",
      "0.586 (+/-0.041) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 500, 'n_estimators': 100}\n",
      "0.592 (+/-0.049) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 1000, 'n_estimators': 10}\n",
      "0.585 (+/-0.041) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 1000, 'n_estimators': 50}\n",
      "0.586 (+/-0.040) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 1000, 'n_estimators': 100}\n",
      "0.590 (+/-0.047) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 2000, 'n_estimators': 10}\n",
      "0.587 (+/-0.043) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 2000, 'n_estimators': 50}\n",
      "0.586 (+/-0.041) for {'max_depth': 20, 'min_samples_leaf': 4, 'min_samples_split': 2000, 'n_estimators': 100}\n",
      "The hyperparameter tuning took:  121.69792604446411 seconds\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import classification_report\n",
    "import time\n",
    "start_time = time.time()\n",
    "parameters = {\n",
    "              \"n_estimators\" : [10,50,100],\n",
    "              'max_depth': [20],\n",
    "              \"min_samples_leaf\" : [4],\n",
    "               'min_samples_split' : [200,500,1000,2000]}\n",
    "\n",
    "\n",
    "opt_classifier = GridSearchCV(RandomForestClassifier(), parameters, cv = Kfold)\n",
    "opt_classifier.fit(X1,y)\n",
    "\n",
    "print(\"Best set of parameters found:\")\n",
    "print(opt_classifier.best_params_)\n",
    "print()\n",
    "print(\"Scores on training set:\")\n",
    "means = opt_classifier.cv_results_['mean_test_score']\n",
    "stds = opt_classifier.cv_results_['std_test_score']\n",
    "for mean, std, params in zip(means, stds, opt_classifier.cv_results_['params']):\n",
    "    print(\"%0.3f (+/-%0.03f) for %r\"\n",
    "          % (mean, std * 2, params))\n",
    "\n",
    "print(\"The hyperparameter tuning took:  %s seconds\" % (time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.naive_bayes import MultinomialNB\n",
    "nb_classifier = MultinomialNB(alpha=1.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the Random forest classifier with 10 fold cross-validation is 0.709018430367\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.006994283199310303"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "start_NB = time.time()\n",
    "\n",
    "NB_scores = cross_val_score(nb_classifier, X1,y, cv=10, scoring='accuracy')\n",
    "print('The accuracy of the Random forest classifier with 10 fold cross-validation is ' + str(NB_scores.mean()))\n",
    "\n",
    "end_NB =time.time()\n",
    "NB_time = (end_NB-start_NB)/60\n",
    "NB_time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the Random forest classifier with 10 fold cross-validation is 0.653895274361\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.006357534726460775"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "start_NB = time.time()\n",
    "\n",
    "NB_scores = cross_val_score(nb_classifier, X2,y, cv=10, scoring='accuracy')\n",
    "print('The accuracy of the Random forest classifier with 10 fold cross-validation is ' + str(NB_scores.mean()))\n",
    "\n",
    "end_NB =time.time()\n",
    "NB_time = (end_NB-start_NB)/60\n",
    "NB_time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import classification_report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best set of parameters found:\n",
      "{'alpha': 1.0}\n",
      "\n",
      "Scores on training set:\n",
      "0.684 (+/-0.048) for {'alpha': 0.01}\n",
      "0.692 (+/-0.056) for {'alpha': 0.1}\n",
      "0.709 (+/-0.045) for {'alpha': 1.0}\n",
      "0.698 (+/-0.038) for {'alpha': 2.0}\n",
      "\n",
      "The hyperparameter tuning took:  1.9929265975952148 seconds\n"
     ]
    }
   ],
   "source": [
    "start_time = time.time()\n",
    "\n",
    "NB = MultinomialNB()\n",
    "\n",
    "param_grid = [{'alpha': [0.01,0.1,1.0,2.0]}]\n",
    "\n",
    "grid = GridSearchCV(NB, param_grid, cv=10, scoring='accuracy')\n",
    "grid.fit(X1,y)\n",
    "\n",
    "print(\"Best set of parameters found:\")\n",
    "print(grid.best_params_)\n",
    "print()\n",
    "\n",
    "print(\"Scores on training set:\")\n",
    "means = grid.cv_results_['mean_test_score']\n",
    "stds = grid.cv_results_['std_test_score']\n",
    "for mean, std, params in zip(means, stds, grid.cv_results_['params']):\n",
    "    print(\"%0.3f (+/-%0.03f) for %r\"\n",
    "          % (mean, std * 2, params))\n",
    "print()\n",
    "print(\"The hyperparameter tuning took:  %s seconds\" % (time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score \n",
    "NB_accuracy = (accuracy_score(nb_classifier_pred ,y_test))*100\n",
    "NB_accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "classifiers = { }\n",
    "classifiers['Classifier'] = ['Logistic Regeression', 'Random Forest', 'Naive Bayesian']\n",
    "classifiers['Run_time'] = [logreg_time, forest_time,NB_time]\n",
    "classifiers['Accuracy'] = [logreg_accuracy, forest_accuracy, NB_accuracy]\n",
    "classifiers = pd.DataFrame(classifiers)\n",
    "classifiers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sns.barplot(y =\"Accuracy\", x=\"Classifier\", data=classifiers); plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "gb = GradientBoostingClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.random.seed(42)\n",
    "# start = time.time()\n",
    "# set ranges of parameters can be selected as different combination of model settings \n",
    "param_dist = {\n",
    "              \"n_estimators\" : [10,20,30,40,50,60,70],\n",
    "              'max_depth': [2, 3, 4,8,10,20],\n",
    "              'bootstrap': [True, False],\n",
    "              'max_features': ['sqrt', 'log2'],   \n",
    "              \"min_samples_leaf\" : [2,4,8],\n",
    "              'criterion': ['gini', 'entropy']}\n",
    "\n",
    "# conduct grid search for different models, same it as model named \"cv_rf\"\n",
    "# GridSearchCV: Exhaustive search over specified parameter values for an estimator.\n",
    "# fit_rf (default settings of RandomForestClassifier) are default settings (baseline model) that set to optimize\n",
    "# cv as corss validation \n",
    "# n_jobs: Number of jobs to run in parallel, default as 1\n",
    "# http://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html\n",
    "\n",
    "cv_rf = GridSearchCV(fit_rf, cv = 5,\n",
    "                     param_grid=param_dist,\n",
    "                     n_jobs = 3)\n",
    "\n",
    "# use model to fit in X_train and y_train data. y_train as output \n",
    "cv_rf.fit(training_set, class_set)\n",
    "\n",
    "print('Best Parameters using grid search: \\n',\n",
    "      cv_rf.best_params_)\n",
    "\n",
    "# end = time.time()\n",
    "# print('Time taken in grid search: {0: .2f}'.format(end - start))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "knn = KNeighborsClassifier(n_neighbors=52)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the knn classifier with 10 fold cross-validation is 0.562303691977\n"
     ]
    }
   ],
   "source": [
    "knn_scores = cross_val_score(knn,X1,y, cv=10, scoring='accuracy')\n",
    "print('The accuracy of the knn classifier with 10 fold cross-validation is ' + str(knn_scores.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the knn classifier with 10 fold cross-validation is 0.64965908694\n"
     ]
    }
   ],
   "source": [
    "knn_scores = cross_val_score(knn,X2,y, cv=10, scoring='accuracy')\n",
    "print('The accuracy of the knn classifier with 10 fold cross-validation is ' + str(knn_scores.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Knn Mean score :  0.635859680968\n",
      "Knn Score variance :  0.000538519767971\n"
     ]
    }
   ],
   "source": [
    "knn_cv_mean = np.mean(knn_scores)\n",
    "knn_cv_variance = np.var(knn_scores)\n",
    "\n",
    "print('Knn Mean score : ', knn_cv_mean)\n",
    "print('Knn Score variance : ', knn_cv_variance)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The optimal number of neighbors is 52.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFXCAYAAAC7nNf0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEELkie/OAGakAkM3MAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28a965744e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "neighbors = range(1,100)\n",
    "\n",
    "# empty list that will hold cv scores\n",
    "cv_scores = []\n",
    "\n",
    "# perform 10-fold cross validation\n",
    "#for k in neighbors:\n",
    "for k in neighbors:\n",
    "    knn = KNeighborsClassifier(n_neighbors=k)\n",
    "    scores = cross_val_score(knn, X2,y, cv=10, 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 = neighbors[MSE.index(min(MSE))]\n",
    "print('The optimal number of neighbors is %d.' % optimal_k)\n",
    "\n",
    "# plot misclassification error vs k \n",
    "plt.plot(neighbors, MSE)\n",
    "plt.xlabel('Number of Neighbors K')\n",
    "plt.ylabel('Misclassification Error')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the SVM classifier with 10 fold cross-validation is 0.727460442556\n"
     ]
    }
   ],
   "source": [
    "from sklearn.svm import SVC\n",
    "svc=SVC(kernel='linear')\n",
    "svc_scores = cross_val_score(svc,z, y_train.values.ravel(), cv=10, scoring='accuracy')\n",
    "print('The accuracy of the SVM classifier with 10 fold cross-validation is ' + str(svc_scores.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 355,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the SVM classifier with 10 fold cross-validation is 0.726293285403\n"
     ]
    }
   ],
   "source": [
    "svc_2=SVC(kernel='rbf',gamma =0.05,C = 10)\n",
    "svc2_scores = cross_val_score(svc,z, y_train.values.ravel(), cv=10, scoring='accuracy')\n",
    "print('The accuracy of the SVM classifier with 10 fold cross-validation is ' + str(svc2_scores.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# K-fold cross validation\n",
    "# preparing the K-fold parameters\n",
    "from sklearn import cross_validation\n",
    "from sklearn.cross_validation import KFold, cross_val_score\n",
    "\n",
    "Kfold = KFold(len(X_train),n_folds = 10,shuffle = True)\n",
    "#print(\"K-fold Cross Val score using SVM is:  %s\" %cross_val_score(SVC(),X,y,cv=Kfold).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 354,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best set of parameters found:\n",
      "{'C': 10, 'gamma': 0.05, 'kernel': 'rbf'}\n",
      "\n",
      "Scores on training set:\n",
      "0.627 (+/-0.033) for {'C': 0.001, 'kernel': 'linear'}\n",
      "0.640 (+/-0.035) for {'C': 0.01, 'kernel': 'linear'}\n",
      "0.724 (+/-0.037) for {'C': 0.1, 'kernel': 'linear'}\n",
      "0.729 (+/-0.029) for {'C': 1, 'kernel': 'linear'}\n",
      "0.718 (+/-0.030) for {'C': 2, 'kernel': 'linear'}\n",
      "0.686 (+/-0.025) for {'C': 10, 'kernel': 'linear'}\n",
      "0.627 (+/-0.033) for {'C': 0.001, 'gamma': 0.005, 'kernel': 'rbf'}\n",
      "0.627 (+/-0.033) for {'C': 0.001, 'gamma': 0.05, 'kernel': 'rbf'}\n",
      "0.627 (+/-0.033) for {'C': 0.001, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.627 (+/-0.033) for {'C': 0.001, 'gamma': 5, 'kernel': 'rbf'}\n",
      "0.627 (+/-0.033) for {'C': 0.01, 'gamma': 0.005, 'kernel': 'rbf'}\n",
      "0.627 (+/-0.033) for {'C': 0.01, 'gamma': 0.05, 'kernel': 'rbf'}\n",
      "0.627 (+/-0.033) for {'C': 0.01, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.627 (+/-0.033) for {'C': 0.01, 'gamma': 5, 'kernel': 'rbf'}\n",
      "0.628 (+/-0.033) for {'C': 0.5, 'gamma': 0.005, 'kernel': 'rbf'}\n",
      "0.674 (+/-0.038) for {'C': 0.5, 'gamma': 0.05, 'kernel': 'rbf'}\n",
      "0.672 (+/-0.030) for {'C': 0.5, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.648 (+/-0.032) for {'C': 0.5, 'gamma': 5, 'kernel': 'rbf'}\n",
      "0.642 (+/-0.036) for {'C': 1, 'gamma': 0.005, 'kernel': 'rbf'}\n",
      "0.731 (+/-0.036) for {'C': 1, 'gamma': 0.05, 'kernel': 'rbf'}\n",
      "0.684 (+/-0.028) for {'C': 1, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.657 (+/-0.032) for {'C': 1, 'gamma': 5, 'kernel': 'rbf'}\n",
      "0.738 (+/-0.038) for {'C': 10, 'gamma': 0.005, 'kernel': 'rbf'}\n",
      "0.741 (+/-0.031) for {'C': 10, 'gamma': 0.05, 'kernel': 'rbf'}\n",
      "0.683 (+/-0.030) for {'C': 10, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.657 (+/-0.032) for {'C': 10, 'gamma': 5, 'kernel': 'rbf'}\n",
      "The hyperparameter tuning took:  3039.569837808609 seconds\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import classification_report\n",
    "import time\n",
    "start_time = time.time()\n",
    "parameters = [{'kernel': ['linear'], 'C': [0.001, 0.01, 0.1, 1, 2, 10]},\n",
    "              {'kernel': ['rbf'],\n",
    "               'gamma': [0.005, 0.05, 0.5, 5],\n",
    "                'C': [0.001, 0.01, 0.5, 1, 10]}]\n",
    "\n",
    "opt_classifier = GridSearchCV(SVC(), parameters, cv = Kfold)\n",
    "opt_classifier.fit(z, y_train.values.ravel())\n",
    "\n",
    "print(\"Best set of parameters found:\")\n",
    "print(opt_classifier.best_params_)\n",
    "print()\n",
    "print(\"Scores on training set:\")\n",
    "means = opt_classifier.cv_results_['mean_test_score']\n",
    "stds = opt_classifier.cv_results_['std_test_score']\n",
    "for mean, std, params in zip(means, stds, opt_classifier.cv_results_['params']):\n",
    "    print(\"%0.3f (+/-%0.03f) for %r\"\n",
    "          % (mean, std * 2, params))\n",
    "\n",
    "print(\"The hyperparameter tuning took:  %s seconds\" % (time.time() - start_time))"
   ]
  }
 ],
 "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}