Collocations.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Collocations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#load all libraries\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import re\n",
    "import nltk\n",
    "from nltk.corpus import stopwords\n",
    "import spacy\n",
    "import string"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data: https://www.kaggle.com/datafiniti/hotel-reviews/data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load reviews data\n",
    "reviews = pd.read_csv('/Users/Nicha/Downloads/hotel_reviews.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 360,
   "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>address</th>\n",
       "      <th>categories</th>\n",
       "      <th>city</th>\n",
       "      <th>country</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>name</th>\n",
       "      <th>postalCode</th>\n",
       "      <th>province</th>\n",
       "      <th>reviews.date</th>\n",
       "      <th>reviews.dateAdded</th>\n",
       "      <th>reviews.doRecommend</th>\n",
       "      <th>reviews.id</th>\n",
       "      <th>reviews.rating</th>\n",
       "      <th>reviews.text</th>\n",
       "      <th>reviews.title</th>\n",
       "      <th>reviews.userCity</th>\n",
       "      <th>reviews.username</th>\n",
       "      <th>reviews.userProvince</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Riviera San Nicol 11/a</td>\n",
       "      <td>Hotels</td>\n",
       "      <td>Mableton</td>\n",
       "      <td>US</td>\n",
       "      <td>45.421611</td>\n",
       "      <td>12.376187</td>\n",
       "      <td>Hotel Russo Palace</td>\n",
       "      <td>30126</td>\n",
       "      <td>GA</td>\n",
       "      <td>2013-09-22T00:00:00Z</td>\n",
       "      <td>2016-10-24T00:00:25Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>Pleasant 10 min walk along the sea front to th...</td>\n",
       "      <td>Good location away from the crouds</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Russ (kent)</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Riviera San Nicol 11/a</td>\n",
       "      <td>Hotels</td>\n",
       "      <td>Mableton</td>\n",
       "      <td>US</td>\n",
       "      <td>45.421611</td>\n",
       "      <td>12.376187</td>\n",
       "      <td>Hotel Russo Palace</td>\n",
       "      <td>30126</td>\n",
       "      <td>GA</td>\n",
       "      <td>2015-04-03T00:00:00Z</td>\n",
       "      <td>2016-10-24T00:00:25Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>Really lovely hotel. Stayed on the very top fl...</td>\n",
       "      <td>Great hotel with Jacuzzi bath!</td>\n",
       "      <td>NaN</td>\n",
       "      <td>A Traveler</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  address categories      city country   latitude  longitude  \\\n",
       "0  Riviera San Nicol 11/a     Hotels  Mableton      US  45.421611  12.376187   \n",
       "1  Riviera San Nicol 11/a     Hotels  Mableton      US  45.421611  12.376187   \n",
       "\n",
       "                 name postalCode province          reviews.date  \\\n",
       "0  Hotel Russo Palace      30126       GA  2013-09-22T00:00:00Z   \n",
       "1  Hotel Russo Palace      30126       GA  2015-04-03T00:00:00Z   \n",
       "\n",
       "      reviews.dateAdded  reviews.doRecommend  reviews.id  reviews.rating  \\\n",
       "0  2016-10-24T00:00:25Z                  NaN         NaN             4.0   \n",
       "1  2016-10-24T00:00:25Z                  NaN         NaN             5.0   \n",
       "\n",
       "                                        reviews.text  \\\n",
       "0  Pleasant 10 min walk along the sea front to th...   \n",
       "1  Really lovely hotel. Stayed on the very top fl...   \n",
       "\n",
       "                        reviews.title reviews.userCity reviews.username  \\\n",
       "0  Good location away from the crouds              NaN      Russ (kent)   \n",
       "1      Great hotel with Jacuzzi bath!              NaN       A Traveler   \n",
       "\n",
       "  reviews.userProvince  \n",
       "0                  NaN  \n",
       "1                  NaN  "
      ]
     },
     "execution_count": 360,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reviews.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Extract only the reviews..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "comments = reviews['reviews.text']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#function to remove non-ascii characters\n",
    "def _removeNonAscii(s): return \"\".join(i for i in s if ord(i)<128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "comments = comments.astype('str')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [],
   "source": [
    "#remove non-ascii characters\n",
    "comments = comments.map(lambda x: _removeNonAscii(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#get stop words of all languages\n",
    "STOPWORDS_DICT = {lang: set(nltk.corpus.stopwords.words(lang)) for lang in nltk.corpus.stopwords.fileids()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#function to detect language based on # of stop words for particular language\n",
    "def get_language(text):\n",
    "    words = set(nltk.wordpunct_tokenize(text.lower()))\n",
    "    lang = max(((lang, len(words & stopwords)) for lang, stopwords in STOPWORDS_DICT.items()), key = lambda x: x[1])[0]\n",
    "    if lang == 'english':\n",
    "        return True\n",
    "    else:\n",
    "        return False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#filter for only english comments\n",
    "eng_comments=comments[comments.apply(get_language)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    Pleasant 10 min walk along the sea front to th...\n",
       "1    Really lovely hotel. Stayed on the very top fl...\n",
       "3    We stayed here for four nights in October. The...\n",
       "4    We stayed here for four nights in October. The...\n",
       "5    We loved staying on the island of Lido! You ne...\n",
       "Name: reviews.text, dtype: object"
      ]
     },
     "execution_count": 234,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eng_comments.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#drop duplicates\n",
    "eng_comments.drop_duplicates(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#load spacy\n",
    "nlp = spacy.load('en')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#function to clean and lemmatize comments\n",
    "def clean_comments(text):\n",
    "    #remove punctuations\n",
    "    regex = re.compile('[' + re.escape(string.punctuation) + '\\\\r\\\\t\\\\n]')\n",
    "    nopunct = regex.sub(\" \", str(text))\n",
    "    #use spacy to lemmatize comments\n",
    "    doc = nlp(nopunct, disable=['parser','ner'])\n",
    "    lemma = [token.lemma_ for token in doc]\n",
    "    return lemma"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#apply function to clean and lemmatize comments\n",
    "lemmatized = eng_comments.map(clean_comments)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "metadata": {},
   "outputs": [],
   "source": [
    "#make sure to lowercase everything\n",
    "lemmatized = lemmatized.map(lambda x: [word.lower() for word in x])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    [pleasant, 10, min, walk, along, the, sea, fro...\n",
       "1    [really, lovely, hotel,  , stay, on, the, very...\n",
       "3    [-pron-, stay, here, for, four, night, in, oct...\n",
       "5    [-pron-, love, stay, on, the, island, of, lido...\n",
       "6    [lovely, view, out, onto, the, lagoon,  , exce...\n",
       "Name: reviews.text, dtype: object"
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lemmatized.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#turn all comments' tokens into one single list\n",
    "unlist_comments = [item for items in lemmatized for item in items]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize NLTK's Bigrams/Trigrams Finder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "bigrams = nltk.collocations.BigramAssocMeasures()\n",
    "trigrams = nltk.collocations.TrigramAssocMeasures()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "bigramFinder = nltk.collocations.BigramCollocationFinder.from_words(unlist_comments)\n",
    "trigramFinder = nltk.collocations.TrigramCollocationFinder.from_words(unlist_comments)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Counting Frequencies of Adjacent Words\n",
    "- Main idea: simply order by frequency\n",
    "- Issues: too sensitive to very frequent pairs and pronouns/articles/prepositions come up often\n",
    "- Solution: filter for only adjectives and nouns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [],
   "source": [
    "bigram_freq = bigramFinder.ngram_fd.items()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {},
   "outputs": [],
   "source": [
    "bigramFreqTable = pd.DataFrame(list(bigram_freq), columns=['bigram','freq']).sort_values(by='freq', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "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>bigram</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>( , -pron-)</td>\n",
       "      <td>29078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>( , the)</td>\n",
       "      <td>21918</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(-pron-, be)</td>\n",
       "      <td>18353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(the, room)</td>\n",
       "      <td>8898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(-pron-, have)</td>\n",
       "      <td>8377</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           bigram   freq\n",
       "0     ( , -pron-)  29078\n",
       "1        ( , the)  21918\n",
       "2    (-pron-, be)  18353\n",
       "3     (the, room)   8898\n",
       "4  (-pron-, have)   8377"
      ]
     },
     "execution_count": 249,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigramFreqTable.head().reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 250,
   "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>bigram</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>152</th>\n",
       "      <td>( , -pron-)</td>\n",
       "      <td>29078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>( , the)</td>\n",
       "      <td>21918</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>(-pron-, be)</td>\n",
       "      <td>18353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>(the, room)</td>\n",
       "      <td>8898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>238</th>\n",
       "      <td>(-pron-, have)</td>\n",
       "      <td>8377</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>(room, be)</td>\n",
       "      <td>8300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>323</th>\n",
       "      <td>(in, the)</td>\n",
       "      <td>8150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190</th>\n",
       "      <td>(be, very)</td>\n",
       "      <td>7708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>248</th>\n",
       "      <td>(be, a)</td>\n",
       "      <td>7263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>237</th>\n",
       "      <td>(and, -pron-)</td>\n",
       "      <td>7012</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             bigram   freq\n",
       "152     ( , -pron-)  29078\n",
       "93         ( , the)  21918\n",
       "58     (-pron-, be)  18353\n",
       "108     (the, room)   8898\n",
       "238  (-pron-, have)   8377\n",
       "109      (room, be)   8300\n",
       "323       (in, the)   8150\n",
       "190      (be, very)   7708\n",
       "248         (be, a)   7263\n",
       "237   (and, -pron-)   7012"
      ]
     },
     "execution_count": 250,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigramFreqTable[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "metadata": {},
   "outputs": [],
   "source": [
    "#get english stopwords\n",
    "en_stopwords = set(stopwords.words('english'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "metadata": {},
   "outputs": [],
   "source": [
    "#function to filter for ADJ/NN bigrams\n",
    "def rightTypes(ngram):\n",
    "    if '-pron-' in ngram or '' in ngram or ' 'in ngram or 't' in ngram:\n",
    "        return False\n",
    "    for word in ngram:\n",
    "        if word in en_stopwords:\n",
    "            return False\n",
    "    acceptable_types = ('JJ', 'JJR', 'JJS', 'NN', 'NNS', 'NNP', 'NNPS')\n",
    "    second_type = ('NN', 'NNS', 'NNP', 'NNPS')\n",
    "    tags = nltk.pos_tag(ngram)\n",
    "    if tags[0][1] in acceptable_types and tags[1][1] in second_type:\n",
    "        return True\n",
    "    else:\n",
    "        return False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "metadata": {},
   "outputs": [],
   "source": [
    "#filter bigrams\n",
    "filtered_bi = bigramFreqTable[bigramFreqTable.bigram.map(lambda x: rightTypes(x))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "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>bigram</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1087</th>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>2674</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>(great, location)</td>\n",
       "      <td>797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>(friendly, staff)</td>\n",
       "      <td>775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5159</th>\n",
       "      <td>(hot, tub)</td>\n",
       "      <td>635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4731</th>\n",
       "      <td>(clean, room)</td>\n",
       "      <td>626</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>(hotel, staff)</td>\n",
       "      <td>539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3098</th>\n",
       "      <td>(continental, breakfast)</td>\n",
       "      <td>531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>(nice, hotel)</td>\n",
       "      <td>530</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4077</th>\n",
       "      <td>(free, breakfast)</td>\n",
       "      <td>522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1905</th>\n",
       "      <td>(great, place)</td>\n",
       "      <td>514</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        bigram  freq\n",
       "1087             (front, desk)  2674\n",
       "73           (great, location)   797\n",
       "270          (friendly, staff)   775\n",
       "5159                (hot, tub)   635\n",
       "4731             (clean, room)   626\n",
       "95              (hotel, staff)   539\n",
       "3098  (continental, breakfast)   531\n",
       "266              (nice, hotel)   530\n",
       "4077         (free, breakfast)   522\n",
       "1905            (great, place)   514"
      ]
     },
     "execution_count": 255,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_bi[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigram_freq = trigramFinder.ngram_fd.items()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramFreqTable = pd.DataFrame(list(trigram_freq), columns=['trigram','freq']).sort_values(by='freq', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "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>trigram</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>( , -pron-, be)</td>\n",
       "      <td>6267</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(the, room, be)</td>\n",
       "      <td>4411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>( , the, room)</td>\n",
       "      <td>3349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>( , -pron-, have)</td>\n",
       "      <td>2681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(the, staff, be)</td>\n",
       "      <td>2641</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             trigram  freq\n",
       "0    ( , -pron-, be)  6267\n",
       "1    (the, room, be)  4411\n",
       "2     ( , the, room)  3349\n",
       "3  ( , -pron-, have)  2681\n",
       "4   (the, staff, be)  2641"
      ]
     },
     "execution_count": 260,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trigramFreqTable.head().reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "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>trigram</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2575</th>\n",
       "      <td>( , -pron-, be)</td>\n",
       "      <td>6267</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114</th>\n",
       "      <td>(the, room, be)</td>\n",
       "      <td>4411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113</th>\n",
       "      <td>( , the, room)</td>\n",
       "      <td>3349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1005</th>\n",
       "      <td>( , -pron-, have)</td>\n",
       "      <td>2681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1455</th>\n",
       "      <td>(the, staff, be)</td>\n",
       "      <td>2641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>682</th>\n",
       "      <td>(the, hotel, be)</td>\n",
       "      <td>2323</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2598</th>\n",
       "      <td>( , there, be)</td>\n",
       "      <td>2181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>666</th>\n",
       "      <td>( , the, staff)</td>\n",
       "      <td>1928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>(-pron-, have, a)</td>\n",
       "      <td>1835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1320</th>\n",
       "      <td>(the, front, desk)</td>\n",
       "      <td>1826</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 trigram  freq\n",
       "2575     ( , -pron-, be)  6267\n",
       "114      (the, room, be)  4411\n",
       "113       ( , the, room)  3349\n",
       "1005   ( , -pron-, have)  2681\n",
       "1455    (the, staff, be)  2641\n",
       "682     (the, hotel, be)  2323\n",
       "2598      ( , there, be)  2181\n",
       "666      ( , the, staff)  1928\n",
       "266    (-pron-, have, a)  1835\n",
       "1320  (the, front, desk)  1826"
      ]
     },
     "execution_count": 261,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trigramFreqTable[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 333,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def rightTypesTri(ngram):\n",
    "    if '-pron-' in ngram or '' in ngram or ' 'in ngram or '  ' in ngram or 't' in ngram:\n",
    "        return False\n",
    "    for word in ngram:\n",
    "        if word in en_stopwords:\n",
    "            return False\n",
    "    first_type = ('JJ', 'JJR', 'JJS', 'NN', 'NNS', 'NNP', 'NNPS')\n",
    "    third_type = ('JJ', 'JJR', 'JJS', 'NN', 'NNS', 'NNP', 'NNPS')\n",
    "    tags = nltk.pos_tag(ngram)\n",
    "    if tags[0][1] in first_type and tags[2][1] in third_type:\n",
    "        return True\n",
    "    else:\n",
    "        return False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_tri = trigramFreqTable[trigramFreqTable.trigram.map(lambda x: rightTypesTri(x))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "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>trigram</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3374</th>\n",
       "      <td>(front, desk, staff)</td>\n",
       "      <td>384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24548</th>\n",
       "      <td>(non, smoking, room)</td>\n",
       "      <td>213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34521</th>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>136</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10218</th>\n",
       "      <td>(front, desk, clerk)</td>\n",
       "      <td>122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12301</th>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37533</th>\n",
       "      <td>(smell, like, smoke)</td>\n",
       "      <td>72</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141458</th>\n",
       "      <td>(old, town, alexandria)</td>\n",
       "      <td>69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18378</th>\n",
       "      <td>(front, desk, person)</td>\n",
       "      <td>65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6088</th>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56372</th>\n",
       "      <td>(great, customer, service)</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           trigram  freq\n",
       "3374          (front, desk, staff)   384\n",
       "24548         (non, smoking, room)   213\n",
       "34521      (holiday, inn, express)   136\n",
       "10218         (front, desk, clerk)   122\n",
       "12301           (flat, screen, tv)    79\n",
       "37533         (smell, like, smoke)    72\n",
       "141458     (old, town, alexandria)    69\n",
       "18378        (front, desk, person)    65\n",
       "6088                (free, wi, fi)    62\n",
       "56372   (great, customer, service)    54"
      ]
     },
     "execution_count": 264,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_tri[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 339,
   "metadata": {},
   "outputs": [],
   "source": [
    "freq_bi = filtered_bi[:20].bigram.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 341,
   "metadata": {},
   "outputs": [],
   "source": [
    "freq_tri = filtered_tri[:20].trigram.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. PMI"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [],
   "source": [
    "bigramFinder.apply_freq_filter(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [],
   "source": [
    "bigramPMITable = pd.DataFrame(list(bigramFinder.score_ngrams(bigrams.pmi)), columns=['bigram','PMI']).sort_values(by='PMI', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "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>bigram</th>\n",
       "      <th>PMI</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(universal, studios)</td>\n",
       "      <td>15.201284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(howard, johnson)</td>\n",
       "      <td>14.954780</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(cracker, barrel)</td>\n",
       "      <td>14.811260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(santa, barbara)</td>\n",
       "      <td>14.522026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(sub, par)</td>\n",
       "      <td>14.088390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>(santana, row)</td>\n",
       "      <td>14.001559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(e, g)</td>\n",
       "      <td>13.687743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(elk, springs)</td>\n",
       "      <td>13.333635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(times, square)</td>\n",
       "      <td>13.161556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>(ear, plug)</td>\n",
       "      <td>13.094932</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 bigram        PMI\n",
       "0  (universal, studios)  15.201284\n",
       "1     (howard, johnson)  14.954780\n",
       "2     (cracker, barrel)  14.811260\n",
       "3      (santa, barbara)  14.522026\n",
       "4            (sub, par)  14.088390\n",
       "5        (santana, row)  14.001559\n",
       "6                (e, g)  13.687743\n",
       "7        (elk, springs)  13.333635\n",
       "8       (times, square)  13.161556\n",
       "9           (ear, plug)  13.094932"
      ]
     },
     "execution_count": 281,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigramPMITable[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramFinder.apply_freq_filter(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramPMITable = pd.DataFrame(list(trigramFinder.score_ngrams(trigrams.pmi)), columns=['trigram','PMI']).sort_values(by='PMI', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "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>trigram</th>\n",
       "      <th>PMI</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(elk, springs, resort)</td>\n",
       "      <td>23.859277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(zion, national, park)</td>\n",
       "      <td>23.223602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>22.598334</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(hard, boil, egg)</td>\n",
       "      <td>22.117153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>21.635639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>(within, walking, distance)</td>\n",
       "      <td>21.585821</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(red, roof, inn)</td>\n",
       "      <td>21.397206</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(simpson, house, inn)</td>\n",
       "      <td>20.803959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>20.634339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>(slide, glass, door)</td>\n",
       "      <td>20.261822</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       trigram        PMI\n",
       "0       (elk, springs, resort)  23.859277\n",
       "1       (zion, national, park)  23.223602\n",
       "2           (flat, screen, tv)  22.598334\n",
       "3            (hard, boil, egg)  22.117153\n",
       "4      (holiday, inn, express)  21.635639\n",
       "5  (within, walking, distance)  21.585821\n",
       "6             (red, roof, inn)  21.397206\n",
       "7        (simpson, house, inn)  20.803959\n",
       "8               (free, wi, fi)  20.634339\n",
       "9         (slide, glass, door)  20.261822"
      ]
     },
     "execution_count": 285,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trigramPMITable[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 343,
   "metadata": {},
   "outputs": [],
   "source": [
    "pmi_bi = bigramPMITable[:20].bigram.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 344,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pmi_tri = trigramPMITable[:20].trigram.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. t-test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "metadata": {},
   "outputs": [],
   "source": [
    "bigramTtable = pd.DataFrame(list(bigramFinder.score_ngrams(bigrams.student_t)), columns=['bigram','t']).sort_values(by='t', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "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>bigram</th>\n",
       "      <th>t</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>( , -pron-)</td>\n",
       "      <td>112.604884</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>( , the)</td>\n",
       "      <td>92.163600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(-pron-, be)</td>\n",
       "      <td>89.806584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(the, room)</td>\n",
       "      <td>79.436600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(be, very)</td>\n",
       "      <td>78.009909</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         bigram           t\n",
       "0   ( , -pron-)  112.604884\n",
       "1      ( , the)   92.163600\n",
       "2  (-pron-, be)   89.806584\n",
       "3   (the, room)   79.436600\n",
       "4    (be, very)   78.009909"
      ]
     },
     "execution_count": 298,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigramTtable.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "metadata": {},
   "outputs": [],
   "source": [
    "filteredT_bi = bigramTtable[bigramTtable.bigram.map(lambda x: rightTypes(x))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 303,
   "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>bigram</th>\n",
       "      <th>t</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>51.576355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>(great, location)</td>\n",
       "      <td>27.429210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120</th>\n",
       "      <td>(friendly, staff)</td>\n",
       "      <td>26.735061</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>136</th>\n",
       "      <td>(hot, tub)</td>\n",
       "      <td>25.152100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161</th>\n",
       "      <td>(continental, breakfast)</td>\n",
       "      <td>22.920096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172</th>\n",
       "      <td>(free, breakfast)</td>\n",
       "      <td>22.405215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>(great, place)</td>\n",
       "      <td>21.472346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>208</th>\n",
       "      <td>(parking, lot)</td>\n",
       "      <td>20.779445</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>218</th>\n",
       "      <td>(customer, service)</td>\n",
       "      <td>20.483320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226</th>\n",
       "      <td>(desk, staff)</td>\n",
       "      <td>20.214107</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       bigram          t\n",
       "21              (front, desk)  51.576355\n",
       "109         (great, location)  27.429210\n",
       "120         (friendly, staff)  26.735061\n",
       "136                (hot, tub)  25.152100\n",
       "161  (continental, breakfast)  22.920096\n",
       "172         (free, breakfast)  22.405215\n",
       "193            (great, place)  21.472346\n",
       "208            (parking, lot)  20.779445\n",
       "218       (customer, service)  20.483320\n",
       "226             (desk, staff)  20.214107"
      ]
     },
     "execution_count": 303,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filteredT_bi[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 304,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramTtable = pd.DataFrame(list(trigramFinder.score_ngrams(trigrams.student_t)), columns=['trigram','t']).sort_values(by='t', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 305,
   "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>trigram</th>\n",
       "      <th>t</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>( , -pron-, be)</td>\n",
       "      <td>72.501091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(the, room, be)</td>\n",
       "      <td>65.285635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>( , the, room)</td>\n",
       "      <td>55.800759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(the, staff, be)</td>\n",
       "      <td>50.822814</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>( , -pron-, have)</td>\n",
       "      <td>49.513033</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             trigram          t\n",
       "0    ( , -pron-, be)  72.501091\n",
       "1    (the, room, be)  65.285635\n",
       "2     ( , the, room)  55.800759\n",
       "3   (the, staff, be)  50.822814\n",
       "4  ( , -pron-, have)  49.513033"
      ]
     },
     "execution_count": 305,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trigramTtable.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 324,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filteredT_tri = trigramTtable[trigramTtable.trigram.map(lambda x: rightTypesTri(x))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 325,
   "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>trigram</th>\n",
       "      <th>t</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>143</th>\n",
       "      <td>(front, desk, staff)</td>\n",
       "      <td>19.593921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>341</th>\n",
       "      <td>(non, smoking, room)</td>\n",
       "      <td>14.594362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>622</th>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>11.661900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>730</th>\n",
       "      <td>(front, desk, clerk)</td>\n",
       "      <td>11.045156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1252</th>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>8.888193</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1416</th>\n",
       "      <td>(smell, like, smoke)</td>\n",
       "      <td>8.485101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1488</th>\n",
       "      <td>(old, town, alexandria)</td>\n",
       "      <td>8.306598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1627</th>\n",
       "      <td>(front, desk, person)</td>\n",
       "      <td>8.061943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1739</th>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>7.874003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2093</th>\n",
       "      <td>(great, customer, service)</td>\n",
       "      <td>7.347582</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         trigram          t\n",
       "143         (front, desk, staff)  19.593921\n",
       "341         (non, smoking, room)  14.594362\n",
       "622      (holiday, inn, express)  11.661900\n",
       "730         (front, desk, clerk)  11.045156\n",
       "1252          (flat, screen, tv)   8.888193\n",
       "1416        (smell, like, smoke)   8.485101\n",
       "1488     (old, town, alexandria)   8.306598\n",
       "1627       (front, desk, person)   8.061943\n",
       "1739              (free, wi, fi)   7.874003\n",
       "2093  (great, customer, service)   7.347582"
      ]
     },
     "execution_count": 325,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filteredT_tri.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "t_bi = filteredT_bi[:20].bigram.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 346,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "t_tri = filteredT_tri[:20].trigram.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Chi-Square"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "bigramChiTable = pd.DataFrame(list(bigramFinder.score_ngrams(bigrams.chi_sq)), columns=['bigram','chi-sq']).sort_values(by='chi-sq', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "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>bigram</th>\n",
       "      <th>chi-sq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(wi, fi)</td>\n",
       "      <td>1.651813e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(cracker, barrel)</td>\n",
       "      <td>1.322475e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(howard, johnson)</td>\n",
       "      <td>1.206747e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(la, quinta)</td>\n",
       "      <td>1.070882e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>1.027741e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>(universal, studios)</td>\n",
       "      <td>9.041648e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(santa, barbara)</td>\n",
       "      <td>8.704756e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(santana, row)</td>\n",
       "      <td>8.364605e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(   , more)</td>\n",
       "      <td>7.497511e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>(flat, screen)</td>\n",
       "      <td>7.115527e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>(french, quarter)</td>\n",
       "      <td>6.727350e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>(elk, springs)</td>\n",
       "      <td>6.709751e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>(walking, distance)</td>\n",
       "      <td>6.055708e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>(pleasantly, surprised)</td>\n",
       "      <td>5.809729e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>(didn, t)</td>\n",
       "      <td>5.613468e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>(red, roof)</td>\n",
       "      <td>5.408683e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>(non, smoking)</td>\n",
       "      <td>5.357550e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>(times, square)</td>\n",
       "      <td>5.130675e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>(air, conditioning)</td>\n",
       "      <td>4.581204e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>(air, conditioner)</td>\n",
       "      <td>4.457097e+05</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     bigram        chi-sq\n",
       "0                  (wi, fi)  1.651813e+06\n",
       "1         (cracker, barrel)  1.322475e+06\n",
       "2         (howard, johnson)  1.206747e+06\n",
       "3              (la, quinta)  1.070882e+06\n",
       "4             (front, desk)  1.027741e+06\n",
       "5      (universal, studios)  9.041648e+05\n",
       "6          (santa, barbara)  8.704756e+05\n",
       "7            (santana, row)  8.364605e+05\n",
       "8               (   , more)  7.497511e+05\n",
       "9            (flat, screen)  7.115527e+05\n",
       "10        (french, quarter)  6.727350e+05\n",
       "11           (elk, springs)  6.709751e+05\n",
       "12      (walking, distance)  6.055708e+05\n",
       "13  (pleasantly, surprised)  5.809729e+05\n",
       "14                (didn, t)  5.613468e+05\n",
       "15              (red, roof)  5.408683e+05\n",
       "16           (non, smoking)  5.357550e+05\n",
       "17          (times, square)  5.130675e+05\n",
       "18      (air, conditioning)  4.581204e+05\n",
       "19       (air, conditioner)  4.457097e+05"
      ]
     },
     "execution_count": 312,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigramChiTable.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 313,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramChiTable = pd.DataFrame(list(trigramFinder.score_ngrams(trigrams.chi_sq)), columns=['trigram','chi-sq']).sort_values(by='chi-sq', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 315,
   "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>trigram</th>\n",
       "      <th>chi-sq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(within, walking, distance)</td>\n",
       "      <td>6.107819e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(elk, springs, resort)</td>\n",
       "      <td>5.784021e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>5.017370e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>4.431679e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(zion, national, park)</td>\n",
       "      <td>3.036709e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>(red, roof, inn)</td>\n",
       "      <td>1.244392e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(hard, boil, egg)</td>\n",
       "      <td>1.046558e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>1.017867e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(simpson, house, inn)</td>\n",
       "      <td>4.578081e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>(within, walk, distance)</td>\n",
       "      <td>3.555933e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>(slide, glass, door)</td>\n",
       "      <td>2.891715e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>(pleasantly, surprise, by)</td>\n",
       "      <td>2.342836e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>(old, town, alexandria)</td>\n",
       "      <td>2.205424e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>(non, smoking, room)</td>\n",
       "      <td>1.974207e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>(air, conditioning, unit)</td>\n",
       "      <td>1.929919e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>(pull, out, couch)</td>\n",
       "      <td>1.648413e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>(king, size, bed)</td>\n",
       "      <td>1.609043e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>(easy, walking, distance)</td>\n",
       "      <td>1.521552e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>(fire, alarm, go)</td>\n",
       "      <td>1.313240e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>(hilton, garden, inn)</td>\n",
       "      <td>1.198932e+07</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        trigram        chi-sq\n",
       "0   (within, walking, distance)  6.107819e+08\n",
       "1        (elk, springs, resort)  5.784021e+08\n",
       "2            (flat, screen, tv)  5.017370e+08\n",
       "3       (holiday, inn, express)  4.431679e+08\n",
       "4        (zion, national, park)  3.036709e+08\n",
       "5              (red, roof, inn)  1.244392e+08\n",
       "6             (hard, boil, egg)  1.046558e+08\n",
       "7                (free, wi, fi)  1.017867e+08\n",
       "8         (simpson, house, inn)  4.578081e+07\n",
       "9      (within, walk, distance)  3.555933e+07\n",
       "10         (slide, glass, door)  2.891715e+07\n",
       "11   (pleasantly, surprise, by)  2.342836e+07\n",
       "12      (old, town, alexandria)  2.205424e+07\n",
       "13         (non, smoking, room)  1.974207e+07\n",
       "14    (air, conditioning, unit)  1.929919e+07\n",
       "15           (pull, out, couch)  1.648413e+07\n",
       "16            (king, size, bed)  1.609043e+07\n",
       "17    (easy, walking, distance)  1.521552e+07\n",
       "18            (fire, alarm, go)  1.313240e+07\n",
       "19        (hilton, garden, inn)  1.198932e+07"
      ]
     },
     "execution_count": 315,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trigramChiTable.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 347,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "chi_bi = bigramChiTable[:20].bigram.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 348,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "chi_tri = trigramChiTable[:20].trigram.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Likelihood"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 316,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "bigramLikTable = pd.DataFrame(list(bigramFinder.score_ngrams(bigrams.likelihood_ratio)), columns=['bigram','likelihood ratio']).sort_values(by='likelihood ratio', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 318,
   "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>bigram</th>\n",
       "      <th>likelihood ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(   , more)</td>\n",
       "      <td>48448.438687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>32462.675748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>( , -pron-)</td>\n",
       "      <td>31120.235250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(didn, t)</td>\n",
       "      <td>26672.134364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(be, very)</td>\n",
       "      <td>24360.396967</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          bigram  likelihood ratio\n",
       "0    (   , more)      48448.438687\n",
       "1  (front, desk)      32462.675748\n",
       "2    ( , -pron-)      31120.235250\n",
       "3      (didn, t)      26672.134364\n",
       "4     (be, very)      24360.396967"
      ]
     },
     "execution_count": 318,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigramLikTable.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 319,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filteredLik_bi = bigramLikTable[bigramLikTable.bigram.map(lambda x: rightTypes(x))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 321,
   "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>bigram</th>\n",
       "      <th>likelihood ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>32462.675748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>(hot, tub)</td>\n",
       "      <td>7400.417612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>(continental, breakfast)</td>\n",
       "      <td>5283.734146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>(customer, service)</td>\n",
       "      <td>4513.277281</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>(wi, fi)</td>\n",
       "      <td>4395.380815</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>(great, location)</td>\n",
       "      <td>4345.972738</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>105</th>\n",
       "      <td>(walk, distance)</td>\n",
       "      <td>4207.577371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113</th>\n",
       "      <td>(parking, lot)</td>\n",
       "      <td>3992.332408</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124</th>\n",
       "      <td>(friendly, staff)</td>\n",
       "      <td>3679.524918</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>(air, conditioner)</td>\n",
       "      <td>3656.082047</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       bigram  likelihood ratio\n",
       "1               (front, desk)      32462.675748\n",
       "51                 (hot, tub)       7400.417612\n",
       "75   (continental, breakfast)       5283.734146\n",
       "93        (customer, service)       4513.277281\n",
       "99                   (wi, fi)       4395.380815\n",
       "100         (great, location)       4345.972738\n",
       "105          (walk, distance)       4207.577371\n",
       "113            (parking, lot)       3992.332408\n",
       "124         (friendly, staff)       3679.524918\n",
       "125        (air, conditioner)       3656.082047"
      ]
     },
     "execution_count": 321,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filteredLik_bi.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 322,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramLikTable = pd.DataFrame(list(trigramFinder.score_ngrams(trigrams.likelihood_ratio)), columns=['trigram','likelihood ratio']).sort_values(by='likelihood ratio', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 323,
   "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>trigram</th>\n",
       "      <th>likelihood ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(the, room, be)</td>\n",
       "      <td>95028.980291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>( , -pron-, be)</td>\n",
       "      <td>85768.942108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(the, staff, be)</td>\n",
       "      <td>77588.060878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(   , more, -pron-)</td>\n",
       "      <td>77213.832210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>( , -pron-, have)</td>\n",
       "      <td>76439.150648</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               trigram  likelihood ratio\n",
       "0      (the, room, be)      95028.980291\n",
       "1      ( , -pron-, be)      85768.942108\n",
       "2     (the, staff, be)      77588.060878\n",
       "3  (   , more, -pron-)      77213.832210\n",
       "4    ( , -pron-, have)      76439.150648"
      ]
     },
     "execution_count": 323,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trigramLikTable.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "metadata": {},
   "outputs": [],
   "source": [
    "filteredLik_tri = trigramLikTable[trigramLikTable.trigram.map(lambda x: rightTypesTri(x))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "metadata": {
    "scrolled": true
   },
   "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>trigram</th>\n",
       "      <th>likelihood ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>(front, desk, clerk)</td>\n",
       "      <td>53404.085856</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>(front, desk, staff)</td>\n",
       "      <td>51768.199453</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>(front, desk, person)</td>\n",
       "      <td>49662.849781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126</th>\n",
       "      <td>(front, desk, attendant)</td>\n",
       "      <td>49209.297668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131</th>\n",
       "      <td>(front, desk, personnel)</td>\n",
       "      <td>49133.163919</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134</th>\n",
       "      <td>(call, front, desk)</td>\n",
       "      <td>49020.607199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>137</th>\n",
       "      <td>(front, desk, people)</td>\n",
       "      <td>49003.888883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139</th>\n",
       "      <td>(front, desk, guy)</td>\n",
       "      <td>48988.631032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>(front, desk, lady)</td>\n",
       "      <td>48976.873913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165</th>\n",
       "      <td>(hotel, front, desk)</td>\n",
       "      <td>48703.645937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3035</th>\n",
       "      <td>(pool, hot, tub)</td>\n",
       "      <td>11296.518002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3335</th>\n",
       "      <td>(non, smoking, room)</td>\n",
       "      <td>9751.555045</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3578</th>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>8602.350756</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3644</th>\n",
       "      <td>(free, continental, breakfast)</td>\n",
       "      <td>8303.777296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3763</th>\n",
       "      <td>(easy, walking, distance)</td>\n",
       "      <td>7819.133186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3870</th>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>7444.810558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3977</th>\n",
       "      <td>(great, customer, service)</td>\n",
       "      <td>7160.001458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4014</th>\n",
       "      <td>(excellent, customer, service)</td>\n",
       "      <td>7070.427775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4087</th>\n",
       "      <td>(good, customer, service)</td>\n",
       "      <td>6884.763461</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4110</th>\n",
       "      <td>(super, friendly, staff)</td>\n",
       "      <td>6828.410056</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                             trigram  likelihood ratio\n",
       "81              (front, desk, clerk)      53404.085856\n",
       "93              (front, desk, staff)      51768.199453\n",
       "116            (front, desk, person)      49662.849781\n",
       "126         (front, desk, attendant)      49209.297668\n",
       "131         (front, desk, personnel)      49133.163919\n",
       "134              (call, front, desk)      49020.607199\n",
       "137            (front, desk, people)      49003.888883\n",
       "139               (front, desk, guy)      48988.631032\n",
       "141              (front, desk, lady)      48976.873913\n",
       "165             (hotel, front, desk)      48703.645937\n",
       "3035                (pool, hot, tub)      11296.518002\n",
       "3335            (non, smoking, room)       9751.555045\n",
       "3578         (holiday, inn, express)       8602.350756\n",
       "3644  (free, continental, breakfast)       8303.777296\n",
       "3763       (easy, walking, distance)       7819.133186\n",
       "3870                  (free, wi, fi)       7444.810558\n",
       "3977      (great, customer, service)       7160.001458\n",
       "4014  (excellent, customer, service)       7070.427775\n",
       "4087       (good, customer, service)       6884.763461\n",
       "4110        (super, friendly, staff)       6828.410056"
      ]
     },
     "execution_count": 335,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filteredLik_tri.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lik_bi = filteredLik_bi[:20].bigram.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 350,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lik_tri = filteredLik_tri[:20].trigram.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bigram Comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 353,
   "metadata": {},
   "outputs": [],
   "source": [
    "bigramsCompare = pd.DataFrame([freq_bi, pmi_bi, t_bi, chi_bi, lik_bi]).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 355,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "bigramsCompare.columns = ['Frequency With Filter', 'PMI', 'T-test With Filter', 'Chi-Sq Test', 'Likeihood Ratio Test With Filter']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 356,
   "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>Frequency With Filter</th>\n",
       "      <th>PMI</th>\n",
       "      <th>T-test With Filter</th>\n",
       "      <th>Chi-Sq Test</th>\n",
       "      <th>Likeihood Ratio Test With Filter</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>(universal, studios)</td>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>(wi, fi)</td>\n",
       "      <td>(front, desk)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(great, location)</td>\n",
       "      <td>(howard, johnson)</td>\n",
       "      <td>(great, location)</td>\n",
       "      <td>(cracker, barrel)</td>\n",
       "      <td>(hot, tub)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(friendly, staff)</td>\n",
       "      <td>(cracker, barrel)</td>\n",
       "      <td>(friendly, staff)</td>\n",
       "      <td>(howard, johnson)</td>\n",
       "      <td>(continental, breakfast)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(hot, tub)</td>\n",
       "      <td>(santa, barbara)</td>\n",
       "      <td>(hot, tub)</td>\n",
       "      <td>(la, quinta)</td>\n",
       "      <td>(customer, service)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(clean, room)</td>\n",
       "      <td>(sub, par)</td>\n",
       "      <td>(continental, breakfast)</td>\n",
       "      <td>(front, desk)</td>\n",
       "      <td>(wi, fi)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>(hotel, staff)</td>\n",
       "      <td>(santana, row)</td>\n",
       "      <td>(free, breakfast)</td>\n",
       "      <td>(universal, studios)</td>\n",
       "      <td>(great, location)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(continental, breakfast)</td>\n",
       "      <td>(e, g)</td>\n",
       "      <td>(great, place)</td>\n",
       "      <td>(santa, barbara)</td>\n",
       "      <td>(walk, distance)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(nice, hotel)</td>\n",
       "      <td>(elk, springs)</td>\n",
       "      <td>(parking, lot)</td>\n",
       "      <td>(santana, row)</td>\n",
       "      <td>(parking, lot)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(free, breakfast)</td>\n",
       "      <td>(times, square)</td>\n",
       "      <td>(customer, service)</td>\n",
       "      <td>(   , more)</td>\n",
       "      <td>(friendly, staff)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>(great, place)</td>\n",
       "      <td>(ear, plug)</td>\n",
       "      <td>(desk, staff)</td>\n",
       "      <td>(flat, screen)</td>\n",
       "      <td>(air, conditioner)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>(desk, staff)</td>\n",
       "      <td>(la, quinta)</td>\n",
       "      <td>(walk, distance)</td>\n",
       "      <td>(french, quarter)</td>\n",
       "      <td>(holiday, inn)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>(parking, lot)</td>\n",
       "      <td>(fire, pit)</td>\n",
       "      <td>(comfortable, bed)</td>\n",
       "      <td>(elk, springs)</td>\n",
       "      <td>(air, conditioning)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>(customer, service)</td>\n",
       "      <td>(san, clemente)</td>\n",
       "      <td>(nice, hotel)</td>\n",
       "      <td>(walking, distance)</td>\n",
       "      <td>(free, breakfast)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>(comfortable, bed)</td>\n",
       "      <td>(san, francisco)</td>\n",
       "      <td>(clean, room)</td>\n",
       "      <td>(pleasantly, surprised)</td>\n",
       "      <td>(next, door)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>(walk, distance)</td>\n",
       "      <td>(san, diego)</td>\n",
       "      <td>(next, door)</td>\n",
       "      <td>(didn, t)</td>\n",
       "      <td>(desk, clerk)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>(next, door)</td>\n",
       "      <td>(san, pedro)</td>\n",
       "      <td>(hotel, staff)</td>\n",
       "      <td>(red, roof)</td>\n",
       "      <td>(easy, access)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>(pool, area)</td>\n",
       "      <td>(french, quarter)</td>\n",
       "      <td>(pool, area)</td>\n",
       "      <td>(non, smoking)</td>\n",
       "      <td>(hampton, inn)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>(good, location)</td>\n",
       "      <td>(wi, fi)</td>\n",
       "      <td>(smoking, room)</td>\n",
       "      <td>(times, square)</td>\n",
       "      <td>(la, quinta)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>(smoking, room)</td>\n",
       "      <td>(chocolate, chip)</td>\n",
       "      <td>(good, location)</td>\n",
       "      <td>(air, conditioning)</td>\n",
       "      <td>(coffee, maker)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>(great, hotel)</td>\n",
       "      <td>(san, antonio)</td>\n",
       "      <td>(desk, clerk)</td>\n",
       "      <td>(air, conditioner)</td>\n",
       "      <td>(smoking, room)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Frequency With Filter                   PMI        T-test With Filter  \\\n",
       "0              (front, desk)  (universal, studios)             (front, desk)   \n",
       "1          (great, location)     (howard, johnson)         (great, location)   \n",
       "2          (friendly, staff)     (cracker, barrel)         (friendly, staff)   \n",
       "3                 (hot, tub)      (santa, barbara)                (hot, tub)   \n",
       "4              (clean, room)            (sub, par)  (continental, breakfast)   \n",
       "5             (hotel, staff)        (santana, row)         (free, breakfast)   \n",
       "6   (continental, breakfast)                (e, g)            (great, place)   \n",
       "7              (nice, hotel)        (elk, springs)            (parking, lot)   \n",
       "8          (free, breakfast)       (times, square)       (customer, service)   \n",
       "9             (great, place)           (ear, plug)             (desk, staff)   \n",
       "10             (desk, staff)          (la, quinta)          (walk, distance)   \n",
       "11            (parking, lot)           (fire, pit)        (comfortable, bed)   \n",
       "12       (customer, service)       (san, clemente)             (nice, hotel)   \n",
       "13        (comfortable, bed)      (san, francisco)             (clean, room)   \n",
       "14          (walk, distance)          (san, diego)              (next, door)   \n",
       "15              (next, door)          (san, pedro)            (hotel, staff)   \n",
       "16              (pool, area)     (french, quarter)              (pool, area)   \n",
       "17          (good, location)              (wi, fi)           (smoking, room)   \n",
       "18           (smoking, room)     (chocolate, chip)          (good, location)   \n",
       "19            (great, hotel)        (san, antonio)             (desk, clerk)   \n",
       "\n",
       "                Chi-Sq Test Likeihood Ratio Test With Filter  \n",
       "0                  (wi, fi)                    (front, desk)  \n",
       "1         (cracker, barrel)                       (hot, tub)  \n",
       "2         (howard, johnson)         (continental, breakfast)  \n",
       "3              (la, quinta)              (customer, service)  \n",
       "4             (front, desk)                         (wi, fi)  \n",
       "5      (universal, studios)                (great, location)  \n",
       "6          (santa, barbara)                 (walk, distance)  \n",
       "7            (santana, row)                   (parking, lot)  \n",
       "8               (   , more)                (friendly, staff)  \n",
       "9            (flat, screen)               (air, conditioner)  \n",
       "10        (french, quarter)                   (holiday, inn)  \n",
       "11           (elk, springs)              (air, conditioning)  \n",
       "12      (walking, distance)                (free, breakfast)  \n",
       "13  (pleasantly, surprised)                     (next, door)  \n",
       "14                (didn, t)                    (desk, clerk)  \n",
       "15              (red, roof)                   (easy, access)  \n",
       "16           (non, smoking)                   (hampton, inn)  \n",
       "17          (times, square)                     (la, quinta)  \n",
       "18      (air, conditioning)                  (coffee, maker)  \n",
       "19       (air, conditioner)                  (smoking, room)  "
      ]
     },
     "execution_count": 356,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigramsCompare"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Trigram Comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 357,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramsCompare = pd.DataFrame([freq_tri, pmi_tri, t_tri, chi_tri, lik_tri]).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 358,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trigramsCompare.columns = ['Frequency With Filter', 'PMI', 'T-test With Filter', 'Chi-Sq Test', 'Likeihood Ratio Test With Filter']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 359,
   "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>Frequency With Filter</th>\n",
       "      <th>PMI</th>\n",
       "      <th>T-test With Filter</th>\n",
       "      <th>Chi-Sq Test</th>\n",
       "      <th>Likeihood Ratio Test With Filter</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(front, desk, staff)</td>\n",
       "      <td>(elk, springs, resort)</td>\n",
       "      <td>(front, desk, staff)</td>\n",
       "      <td>(within, walking, distance)</td>\n",
       "      <td>(front, desk, clerk)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(non, smoking, room)</td>\n",
       "      <td>(zion, national, park)</td>\n",
       "      <td>(non, smoking, room)</td>\n",
       "      <td>(elk, springs, resort)</td>\n",
       "      <td>(front, desk, staff)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>(front, desk, person)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(front, desk, clerk)</td>\n",
       "      <td>(hard, boil, egg)</td>\n",
       "      <td>(front, desk, clerk)</td>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>(front, desk, attendant)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "      <td>(flat, screen, tv)</td>\n",
       "      <td>(zion, national, park)</td>\n",
       "      <td>(front, desk, personnel)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>(smell, like, smoke)</td>\n",
       "      <td>(within, walking, distance)</td>\n",
       "      <td>(smell, like, smoke)</td>\n",
       "      <td>(red, roof, inn)</td>\n",
       "      <td>(call, front, desk)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(old, town, alexandria)</td>\n",
       "      <td>(red, roof, inn)</td>\n",
       "      <td>(old, town, alexandria)</td>\n",
       "      <td>(hard, boil, egg)</td>\n",
       "      <td>(front, desk, people)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(front, desk, person)</td>\n",
       "      <td>(simpson, house, inn)</td>\n",
       "      <td>(front, desk, person)</td>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>(front, desk, guy)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>(free, wi, fi)</td>\n",
       "      <td>(simpson, house, inn)</td>\n",
       "      <td>(front, desk, lady)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>(great, customer, service)</td>\n",
       "      <td>(slide, glass, door)</td>\n",
       "      <td>(great, customer, service)</td>\n",
       "      <td>(within, walk, distance)</td>\n",
       "      <td>(hotel, front, desk)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>(good, night, sleep)</td>\n",
       "      <td>(air, conditioning, unit)</td>\n",
       "      <td>(good, night, sleep)</td>\n",
       "      <td>(slide, glass, door)</td>\n",
       "      <td>(pool, hot, tub)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>(front, desk, people)</td>\n",
       "      <td>(pleasantly, surprise, by)</td>\n",
       "      <td>(front, desk, people)</td>\n",
       "      <td>(pleasantly, surprise, by)</td>\n",
       "      <td>(non, smoking, room)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>(easy, walking, distance)</td>\n",
       "      <td>(fire, alarm, go)</td>\n",
       "      <td>(easy, walking, distance)</td>\n",
       "      <td>(old, town, alexandria)</td>\n",
       "      <td>(holiday, inn, express)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>(red, roof, inn)</td>\n",
       "      <td>(hilton, garden, inn)</td>\n",
       "      <td>(red, roof, inn)</td>\n",
       "      <td>(non, smoking, room)</td>\n",
       "      <td>(free, continental, breakfast)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>(first, time, stay)</td>\n",
       "      <td>(pull, out, couch)</td>\n",
       "      <td>(first, time, stay)</td>\n",
       "      <td>(air, conditioning, unit)</td>\n",
       "      <td>(easy, walking, distance)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>(second, time, stay)</td>\n",
       "      <td>(conveniently, locate, near)</td>\n",
       "      <td>(second, time, stay)</td>\n",
       "      <td>(pull, out, couch)</td>\n",
       "      <td>(free, wi, fi)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>(elk, springs, resort)</td>\n",
       "      <td>(of, times, square)</td>\n",
       "      <td>(elk, springs, resort)</td>\n",
       "      <td>(king, size, bed)</td>\n",
       "      <td>(great, customer, service)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>(pet, friendly, hotel)</td>\n",
       "      <td>(easy, walking, distance)</td>\n",
       "      <td>(queen, size, bed)</td>\n",
       "      <td>(easy, walking, distance)</td>\n",
       "      <td>(excellent, customer, service)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>(queen, size, bed)</td>\n",
       "      <td>(old, town, alexandria)</td>\n",
       "      <td>(call, front, desk)</td>\n",
       "      <td>(fire, alarm, go)</td>\n",
       "      <td>(good, customer, service)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>(call, front, desk)</td>\n",
       "      <td>(within, walk, distance)</td>\n",
       "      <td>(pet, friendly, hotel)</td>\n",
       "      <td>(hilton, garden, inn)</td>\n",
       "      <td>(super, friendly, staff)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Frequency With Filter                           PMI  \\\n",
       "0         (front, desk, staff)        (elk, springs, resort)   \n",
       "1         (non, smoking, room)        (zion, national, park)   \n",
       "2      (holiday, inn, express)            (flat, screen, tv)   \n",
       "3         (front, desk, clerk)             (hard, boil, egg)   \n",
       "4           (flat, screen, tv)       (holiday, inn, express)   \n",
       "5         (smell, like, smoke)   (within, walking, distance)   \n",
       "6      (old, town, alexandria)              (red, roof, inn)   \n",
       "7        (front, desk, person)         (simpson, house, inn)   \n",
       "8               (free, wi, fi)                (free, wi, fi)   \n",
       "9   (great, customer, service)          (slide, glass, door)   \n",
       "10        (good, night, sleep)     (air, conditioning, unit)   \n",
       "11       (front, desk, people)    (pleasantly, surprise, by)   \n",
       "12   (easy, walking, distance)             (fire, alarm, go)   \n",
       "13            (red, roof, inn)         (hilton, garden, inn)   \n",
       "14         (first, time, stay)            (pull, out, couch)   \n",
       "15        (second, time, stay)  (conveniently, locate, near)   \n",
       "16      (elk, springs, resort)           (of, times, square)   \n",
       "17      (pet, friendly, hotel)     (easy, walking, distance)   \n",
       "18          (queen, size, bed)       (old, town, alexandria)   \n",
       "19         (call, front, desk)      (within, walk, distance)   \n",
       "\n",
       "            T-test With Filter                  Chi-Sq Test  \\\n",
       "0         (front, desk, staff)  (within, walking, distance)   \n",
       "1         (non, smoking, room)       (elk, springs, resort)   \n",
       "2      (holiday, inn, express)           (flat, screen, tv)   \n",
       "3         (front, desk, clerk)      (holiday, inn, express)   \n",
       "4           (flat, screen, tv)       (zion, national, park)   \n",
       "5         (smell, like, smoke)             (red, roof, inn)   \n",
       "6      (old, town, alexandria)            (hard, boil, egg)   \n",
       "7        (front, desk, person)               (free, wi, fi)   \n",
       "8               (free, wi, fi)        (simpson, house, inn)   \n",
       "9   (great, customer, service)     (within, walk, distance)   \n",
       "10        (good, night, sleep)         (slide, glass, door)   \n",
       "11       (front, desk, people)   (pleasantly, surprise, by)   \n",
       "12   (easy, walking, distance)      (old, town, alexandria)   \n",
       "13            (red, roof, inn)         (non, smoking, room)   \n",
       "14         (first, time, stay)    (air, conditioning, unit)   \n",
       "15        (second, time, stay)           (pull, out, couch)   \n",
       "16      (elk, springs, resort)            (king, size, bed)   \n",
       "17          (queen, size, bed)    (easy, walking, distance)   \n",
       "18         (call, front, desk)            (fire, alarm, go)   \n",
       "19      (pet, friendly, hotel)        (hilton, garden, inn)   \n",
       "\n",
       "   Likeihood Ratio Test With Filter  \n",
       "0              (front, desk, clerk)  \n",
       "1              (front, desk, staff)  \n",
       "2             (front, desk, person)  \n",
       "3          (front, desk, attendant)  \n",
       "4          (front, desk, personnel)  \n",
       "5               (call, front, desk)  \n",
       "6             (front, desk, people)  \n",
       "7                (front, desk, guy)  \n",
       "8               (front, desk, lady)  \n",
       "9              (hotel, front, desk)  \n",
       "10                 (pool, hot, tub)  \n",
       "11             (non, smoking, room)  \n",
       "12          (holiday, inn, express)  \n",
       "13   (free, continental, breakfast)  \n",
       "14        (easy, walking, distance)  \n",
       "15                   (free, wi, fi)  \n",
       "16       (great, customer, service)  \n",
       "17   (excellent, customer, service)  \n",
       "18        (good, customer, service)  \n",
       "19         (super, friendly, staff)  "
      ]
     },
     "execution_count": 359,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trigramsCompare"
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}