Project_Sentiment_Analysis/Sentiment_Classification_Projects.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sentiment Classification & How To \"Frame Problems\" for a Neural Network\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Zikri Bayraktar\n",
    "\n",
    "\n",
    "Lecture notes describing how to approach sentiment analysis using movie sentiment classification."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "56bb3cba-260c-4ebe-9ed6-b995b4c72aa3"
    }
   },
   "source": [
    "# Lesson: Curate a Dataset<a id='lesson_1'></a>\n",
    "The cells from here until Project 1 include code Andrew shows in the videos leading up to mini project 1. We've included them so you can run the code along with the videos without having to type in everything."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "nbpresent": {
     "id": "eba2b193-0419-431e-8db9-60f34dd3fe83"
    }
   },
   "outputs": [],
   "source": [
    "def pretty_print_review_and_label(i):\n",
    "    print(labels[i] + \"\\t:\\t\" + reviews[i][:80] + \"...\")\n",
    "\n",
    "g = open('reviews.txt','r') # What we know!\n",
    "reviews = list(map(lambda x:x[:-1],g.readlines()))\n",
    "g.close()\n",
    "\n",
    "g = open('labels.txt','r') # What we WANT to know!\n",
    "labels = list(map(lambda x:x[:-1].upper(),g.readlines()))\n",
    "g.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note:** The data in `reviews.txt` we're using has already been preprocessed a bit and contains only lower case characters. If we were working from raw data, where we didn't know it was all lower case, we would want to add a step here to convert it. That's so we treat different variations of the same word, like `The`, `the`, and `THE`, all the same way."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25000"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(reviews)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "nbpresent": {
     "id": "bb95574b-21a0-4213-ae50-34363cf4f87f"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'bromwell high is a cartoon comedy . it ran at the same time as some other programs about school life  such as  teachers  . my   years in the teaching profession lead me to believe that bromwell high  s satire is much closer to reality than is  teachers  . the scramble to survive financially  the insightful students who can see right through their pathetic teachers  pomp  the pettiness of the whole situation  all remind me of the schools i knew and their students . when i saw the episode in which a student repeatedly tried to burn down the school  i immediately recalled . . . . . . . . . at . . . . . . . . . . high . a classic line inspector i  m here to sack one of your teachers . student welcome to bromwell high . i expect that many adults of my age think that bromwell high is far fetched . what a pity that it isn  t   '"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reviews[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "nbpresent": {
     "id": "e0408810-c424-4ed4-afb9-1735e9ddbd0a"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'POSITIVE'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Lesson: Develop a Predictive Theory<a id='lesson_2'></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "nbpresent": {
     "id": "e67a709f-234f-4493-bae6-4fb192141ee0"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "labels.txt \t : \t reviews.txt\n",
      "\n",
      "NEGATIVE\t:\tthis movie is terrible but it has some good effects .  ...\n",
      "POSITIVE\t:\tadrian pasdar is excellent is this film . he makes a fascinating woman .  ...\n",
      "NEGATIVE\t:\tcomment this movie is impossible . is terrible  very improbable  bad interpretat...\n",
      "POSITIVE\t:\texcellent episode movie ala pulp fiction .  days   suicides . it doesnt get more...\n",
      "NEGATIVE\t:\tif you haven  t seen this  it  s terrible . it is pure trash . i saw this about ...\n",
      "POSITIVE\t:\tthis schiffer guy is a real genius  the movie is of excellent quality and both e...\n"
     ]
    }
   ],
   "source": [
    "print(\"labels.txt \\t : \\t reviews.txt\\n\")\n",
    "pretty_print_review_and_label(2137)\n",
    "pretty_print_review_and_label(12816)\n",
    "pretty_print_review_and_label(6267)\n",
    "pretty_print_review_and_label(21934)\n",
    "pretty_print_review_and_label(5297)\n",
    "pretty_print_review_and_label(4998)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 1: Quick Theory Validation<a id='project_1'></a>\n",
    "\n",
    "There are multiple ways to implement these projects, but in order to get your code closer to what Andrew shows in his solutions, we've provided some hints and starter code throughout this notebook.\n",
    "\n",
    "You'll find the [Counter](https://docs.python.org/2/library/collections.html#collections.Counter) class to be useful in this exercise, as well as the [numpy](https://docs.scipy.org/doc/numpy/reference/) library."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from collections import Counter\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We'll create three `Counter` objects, one for words from postive reviews, one for words from negative reviews, and one for all the words."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Create three Counter objects to store positive, negative and total counts\n",
    "positive_counts = Counter()\n",
    "negative_counts = Counter()\n",
    "total_counts = Counter()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**TODO:** Examine all the reviews. For each word in a positive review, increase the count for that word in both your positive counter and the total words counter; likewise, for each word in a negative review, increase the count for that word in both your negative counter and the total words counter.\n",
    "\n",
    "**Note:** Throughout these projects, you should use `split(' ')` to divide a piece of text (such as a review) into individual words. If you use `split()` instead, you'll get slightly different results than what the videos and solutions show."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# TODO: Loop over all the words in all the reviews and increment the counts in the appropriate counter objects\n",
    "for lb, rw in zip(labels, reviews):\n",
    "    thisReview = rw.split(' ')\n",
    "    if lb == \"POSITIVE\":\n",
    "        for word in thisReview:\n",
    "            positive_counts[word] += 1\n",
    "            total_counts[word] += 1\n",
    "    else:  #lb==\"NEGATIVE\"\n",
    "        for word in thisReview:\n",
    "            negative_counts[word] += 1\n",
    "            total_counts[word] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following two cells to list the words used in positive reviews and negative reviews, respectively, ordered from most to least commonly used. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('', 550468),\n",
       " ('the', 173324),\n",
       " ('.', 159654),\n",
       " ('and', 89722),\n",
       " ('a', 83688),\n",
       " ('of', 76855),\n",
       " ('to', 66746),\n",
       " ('is', 57245),\n",
       " ('in', 50215),\n",
       " ('br', 49235),\n",
       " ('it', 48025),\n",
       " ('i', 40743),\n",
       " ('that', 35630),\n",
       " ('this', 35080),\n",
       " ('s', 33815),\n",
       " ('as', 26308),\n",
       " ('with', 23247),\n",
       " ('for', 22416),\n",
       " ('was', 21917),\n",
       " ('film', 20937),\n",
       " ('but', 20822),\n",
       " ('movie', 19074),\n",
       " ('his', 17227),\n",
       " ('on', 17008),\n",
       " ('you', 16681),\n",
       " ('he', 16282),\n",
       " ('are', 14807),\n",
       " ('not', 14272),\n",
       " ('t', 13720),\n",
       " ('one', 13655),\n",
       " ('have', 12587),\n",
       " ('be', 12416),\n",
       " ('by', 11997),\n",
       " ('all', 11942),\n",
       " ('who', 11464),\n",
       " ('an', 11294),\n",
       " ('at', 11234),\n",
       " ('from', 10767),\n",
       " ('her', 10474),\n",
       " ('they', 9895),\n",
       " ('has', 9186),\n",
       " ('so', 9154),\n",
       " ('like', 9038),\n",
       " ('about', 8313),\n",
       " ('very', 8305),\n",
       " ('out', 8134),\n",
       " ('there', 8057),\n",
       " ('she', 7779),\n",
       " ('what', 7737),\n",
       " ('or', 7732),\n",
       " ('good', 7720),\n",
       " ('more', 7521),\n",
       " ('when', 7456),\n",
       " ('some', 7441),\n",
       " ('if', 7285),\n",
       " ('just', 7152),\n",
       " ('can', 7001),\n",
       " ('story', 6780),\n",
       " ('time', 6515),\n",
       " ('my', 6488),\n",
       " ('great', 6419),\n",
       " ('well', 6405),\n",
       " ('up', 6321),\n",
       " ('which', 6267),\n",
       " ('their', 6107),\n",
       " ('see', 6026),\n",
       " ('also', 5550),\n",
       " ('we', 5531),\n",
       " ('really', 5476),\n",
       " ('would', 5400),\n",
       " ('will', 5218),\n",
       " ('me', 5167),\n",
       " ('had', 5148),\n",
       " ('only', 5137),\n",
       " ('him', 5018),\n",
       " ('even', 4964),\n",
       " ('most', 4864),\n",
       " ('other', 4858),\n",
       " ('were', 4782),\n",
       " ('first', 4755),\n",
       " ('than', 4736),\n",
       " ('much', 4685),\n",
       " ('its', 4622),\n",
       " ('no', 4574),\n",
       " ('into', 4544),\n",
       " ('people', 4479),\n",
       " ('best', 4319),\n",
       " ('love', 4301),\n",
       " ('get', 4272),\n",
       " ('how', 4213),\n",
       " ('life', 4199),\n",
       " ('been', 4189),\n",
       " ('because', 4079),\n",
       " ('way', 4036),\n",
       " ('do', 3941),\n",
       " ('made', 3823),\n",
       " ('films', 3813),\n",
       " ('them', 3805),\n",
       " ('after', 3800),\n",
       " ('many', 3766),\n",
       " ('two', 3733),\n",
       " ('too', 3659),\n",
       " ('think', 3655),\n",
       " ('movies', 3586),\n",
       " ('characters', 3560),\n",
       " ('character', 3514),\n",
       " ('don', 3468),\n",
       " ('man', 3460),\n",
       " ('show', 3432),\n",
       " ('watch', 3424),\n",
       " ('seen', 3414),\n",
       " ('then', 3358),\n",
       " ('little', 3341),\n",
       " ('still', 3340),\n",
       " ('make', 3303),\n",
       " ('could', 3237),\n",
       " ('never', 3226),\n",
       " ('being', 3217),\n",
       " ('where', 3173),\n",
       " ('does', 3069),\n",
       " ('over', 3017),\n",
       " ('any', 3002),\n",
       " ('while', 2899),\n",
       " ('know', 2833),\n",
       " ('did', 2790),\n",
       " ('years', 2758),\n",
       " ('here', 2740),\n",
       " ('ever', 2734),\n",
       " ('end', 2696),\n",
       " ('these', 2694),\n",
       " ('such', 2590),\n",
       " ('real', 2568),\n",
       " ('scene', 2567),\n",
       " ('back', 2547),\n",
       " ('those', 2485),\n",
       " ('though', 2475),\n",
       " ('off', 2463),\n",
       " ('new', 2458),\n",
       " ('your', 2453),\n",
       " ('go', 2440),\n",
       " ('acting', 2437),\n",
       " ('plot', 2432),\n",
       " ('world', 2429),\n",
       " ('scenes', 2427),\n",
       " ('say', 2414),\n",
       " ('through', 2409),\n",
       " ('makes', 2390),\n",
       " ('better', 2381),\n",
       " ('now', 2368),\n",
       " ('work', 2346),\n",
       " ('young', 2343),\n",
       " ('old', 2311),\n",
       " ('ve', 2307),\n",
       " ('find', 2272),\n",
       " ('both', 2248),\n",
       " ('before', 2177),\n",
       " ('us', 2162),\n",
       " ('again', 2158),\n",
       " ('series', 2153),\n",
       " ('quite', 2143),\n",
       " ('something', 2135),\n",
       " ('cast', 2133),\n",
       " ('should', 2121),\n",
       " ('part', 2098),\n",
       " ('always', 2088),\n",
       " ('lot', 2087),\n",
       " ('another', 2075),\n",
       " ('actors', 2047),\n",
       " ('director', 2040),\n",
       " ('family', 2032),\n",
       " ('between', 2016),\n",
       " ('own', 2016),\n",
       " ('m', 1998),\n",
       " ('may', 1997),\n",
       " ('same', 1972),\n",
       " ('role', 1967),\n",
       " ('watching', 1966),\n",
       " ('every', 1954),\n",
       " ('funny', 1953),\n",
       " ('doesn', 1935),\n",
       " ('performance', 1928),\n",
       " ('few', 1918),\n",
       " ('bad', 1907),\n",
       " ('look', 1900),\n",
       " ('re', 1884),\n",
       " ('why', 1855),\n",
       " ('things', 1849),\n",
       " ('times', 1832),\n",
       " ('big', 1815),\n",
       " ('however', 1795),\n",
       " ('actually', 1790),\n",
       " ('action', 1789),\n",
       " ('going', 1783),\n",
       " ('bit', 1757),\n",
       " ('comedy', 1742),\n",
       " ('down', 1740),\n",
       " ('music', 1738),\n",
       " ('must', 1728),\n",
       " ('take', 1709),\n",
       " ('saw', 1692),\n",
       " ('long', 1690),\n",
       " ('right', 1688),\n",
       " ('fun', 1686),\n",
       " ('fact', 1684),\n",
       " ('excellent', 1683),\n",
       " ('around', 1674),\n",
       " ('didn', 1672),\n",
       " ('without', 1671),\n",
       " ('thing', 1662),\n",
       " ('thought', 1639),\n",
       " ('got', 1635),\n",
       " ('each', 1630),\n",
       " ('day', 1614),\n",
       " ('feel', 1597),\n",
       " ('seems', 1596),\n",
       " ('come', 1594),\n",
       " ('done', 1586),\n",
       " ('beautiful', 1580),\n",
       " ('especially', 1572),\n",
       " ('played', 1571),\n",
       " ('almost', 1566),\n",
       " ('want', 1562),\n",
       " ('yet', 1556),\n",
       " ('give', 1553),\n",
       " ('pretty', 1549),\n",
       " ('last', 1543),\n",
       " ('since', 1519),\n",
       " ('different', 1504),\n",
       " ('although', 1501),\n",
       " ('gets', 1490),\n",
       " ('true', 1487),\n",
       " ('interesting', 1481),\n",
       " ('job', 1470),\n",
       " ('enough', 1455),\n",
       " ('our', 1454),\n",
       " ('shows', 1447),\n",
       " ('horror', 1441),\n",
       " ('woman', 1439),\n",
       " ('tv', 1400),\n",
       " ('probably', 1398),\n",
       " ('father', 1395),\n",
       " ('original', 1393),\n",
       " ('girl', 1390),\n",
       " ('point', 1379),\n",
       " ('plays', 1378),\n",
       " ('wonderful', 1372),\n",
       " ('course', 1358),\n",
       " ('far', 1358),\n",
       " ('john', 1350),\n",
       " ('rather', 1340),\n",
       " ('isn', 1328),\n",
       " ('ll', 1326),\n",
       " ('later', 1324),\n",
       " ('dvd', 1324),\n",
       " ('whole', 1310),\n",
       " ('war', 1310),\n",
       " ('d', 1307),\n",
       " ('away', 1306),\n",
       " ('found', 1306),\n",
       " ('screen', 1305),\n",
       " ('nothing', 1300),\n",
       " ('year', 1297),\n",
       " ('once', 1296),\n",
       " ('hard', 1294),\n",
       " ('together', 1280),\n",
       " ('set', 1277),\n",
       " ('am', 1277),\n",
       " ('having', 1266),\n",
       " ('making', 1265),\n",
       " ('place', 1263),\n",
       " ('comes', 1260),\n",
       " ('might', 1260),\n",
       " ('sure', 1253),\n",
       " ('american', 1248),\n",
       " ('play', 1245),\n",
       " ('kind', 1244),\n",
       " ('perfect', 1242),\n",
       " ('takes', 1242),\n",
       " ('performances', 1237),\n",
       " ('himself', 1230),\n",
       " ('worth', 1221),\n",
       " ('everyone', 1221),\n",
       " ('anyone', 1214),\n",
       " ('actor', 1203),\n",
       " ('three', 1201),\n",
       " ('wife', 1196),\n",
       " ('classic', 1192),\n",
       " ('goes', 1186),\n",
       " ('ending', 1178),\n",
       " ('version', 1168),\n",
       " ('star', 1149),\n",
       " ('enjoy', 1146),\n",
       " ('book', 1142),\n",
       " ('nice', 1132),\n",
       " ('everything', 1128),\n",
       " ('during', 1124),\n",
       " ('put', 1118),\n",
       " ('seeing', 1111),\n",
       " ('least', 1102),\n",
       " ('house', 1100),\n",
       " ('high', 1095),\n",
       " ('watched', 1094),\n",
       " ('loved', 1087),\n",
       " ('men', 1087),\n",
       " ('night', 1082),\n",
       " ('anything', 1075),\n",
       " ('believe', 1071),\n",
       " ('guy', 1071),\n",
       " ('top', 1063),\n",
       " ('amazing', 1058),\n",
       " ('hollywood', 1056),\n",
       " ('looking', 1053),\n",
       " ('main', 1044),\n",
       " ('definitely', 1043),\n",
       " ('gives', 1031),\n",
       " ('home', 1029),\n",
       " ('seem', 1028),\n",
       " ('episode', 1023),\n",
       " ('sense', 1020),\n",
       " ('audience', 1020),\n",
       " ('truly', 1017),\n",
       " ('special', 1011),\n",
       " ('second', 1009),\n",
       " ('short', 1009),\n",
       " ('fan', 1009),\n",
       " ('mind', 1005),\n",
       " ('human', 1001),\n",
       " ('recommend', 999),\n",
       " ('full', 996),\n",
       " ('black', 995),\n",
       " ('help', 991),\n",
       " ('along', 989),\n",
       " ('trying', 987),\n",
       " ('small', 986),\n",
       " ('death', 985),\n",
       " ('friends', 981),\n",
       " ('remember', 974),\n",
       " ('often', 970),\n",
       " ('said', 966),\n",
       " ('favorite', 962),\n",
       " ('heart', 959),\n",
       " ('early', 957),\n",
       " ('left', 956),\n",
       " ('until', 955),\n",
       " ('let', 954),\n",
       " ('script', 954),\n",
       " ('maybe', 937),\n",
       " ('today', 936),\n",
       " ('less', 934),\n",
       " ('live', 934),\n",
       " ('moments', 933),\n",
       " ('others', 929),\n",
       " ('brilliant', 926),\n",
       " ('shot', 925),\n",
       " ('liked', 923),\n",
       " ('become', 916),\n",
       " ('won', 915),\n",
       " ('used', 910),\n",
       " ('style', 907),\n",
       " ('mother', 895),\n",
       " ('lives', 894),\n",
       " ('came', 893),\n",
       " ('stars', 890),\n",
       " ('cinema', 889),\n",
       " ('looks', 885),\n",
       " ('perhaps', 884),\n",
       " ('read', 882),\n",
       " ('enjoyed', 879),\n",
       " ('boy', 875),\n",
       " ('drama', 873),\n",
       " ('highly', 871),\n",
       " ('given', 870),\n",
       " ('playing', 867),\n",
       " ('use', 864),\n",
       " ('next', 859),\n",
       " ('women', 858),\n",
       " ('fine', 857),\n",
       " ('effects', 856),\n",
       " ('kids', 854),\n",
       " ('entertaining', 853),\n",
       " ('need', 852),\n",
       " ('line', 850),\n",
       " ('works', 848),\n",
       " ('someone', 847),\n",
       " ('mr', 836),\n",
       " ('simply', 835),\n",
       " ('picture', 833),\n",
       " ('children', 833),\n",
       " ('face', 831),\n",
       " ('friend', 831),\n",
       " ('keep', 831),\n",
       " ('dark', 830),\n",
       " ('certainly', 828),\n",
       " ('overall', 828),\n",
       " ('minutes', 827),\n",
       " ('wasn', 824),\n",
       " ('history', 822),\n",
       " ('finally', 820),\n",
       " ('couple', 816),\n",
       " ('against', 815),\n",
       " ('son', 809),\n",
       " ('understand', 808),\n",
       " ('lost', 807),\n",
       " ('michael', 805),\n",
       " ('else', 801),\n",
       " ('throughout', 798),\n",
       " ('fans', 797),\n",
       " ('city', 792),\n",
       " ('reason', 789),\n",
       " ('production', 787),\n",
       " ('written', 787),\n",
       " ('several', 784),\n",
       " ('school', 783),\n",
       " ('based', 781),\n",
       " ('rest', 781),\n",
       " ('try', 780),\n",
       " ('dead', 776),\n",
       " ('hope', 775),\n",
       " ('strong', 768),\n",
       " ('white', 765),\n",
       " ('tell', 759),\n",
       " ('itself', 758),\n",
       " ('half', 753),\n",
       " ('person', 749),\n",
       " ('sometimes', 746),\n",
       " ('start', 744),\n",
       " ('past', 744),\n",
       " ('genre', 743),\n",
       " ('final', 739),\n",
       " ('beginning', 739),\n",
       " ('town', 738),\n",
       " ('art', 734),\n",
       " ('game', 732),\n",
       " ('humor', 732),\n",
       " ('idea', 731),\n",
       " ('yes', 731),\n",
       " ('late', 730),\n",
       " ('becomes', 729),\n",
       " ('despite', 729),\n",
       " ('able', 726),\n",
       " ('case', 726),\n",
       " ('money', 723),\n",
       " ('completely', 721),\n",
       " ('child', 721),\n",
       " ('side', 719),\n",
       " ('camera', 716),\n",
       " ('getting', 714),\n",
       " ('instead', 712),\n",
       " ('soon', 702),\n",
       " ('under', 700),\n",
       " ('viewer', 699),\n",
       " ('age', 697),\n",
       " ('days', 696),\n",
       " ('stories', 696),\n",
       " ('felt', 694),\n",
       " ('simple', 694),\n",
       " ('roles', 693),\n",
       " ('video', 688),\n",
       " ('either', 683),\n",
       " ('name', 683),\n",
       " ('doing', 677),\n",
       " ('turns', 674),\n",
       " ('close', 671),\n",
       " ('wants', 671),\n",
       " ('title', 669),\n",
       " ('wrong', 668),\n",
       " ('went', 666),\n",
       " ('james', 665),\n",
       " ('evil', 659),\n",
       " ('budget', 657),\n",
       " ('episodes', 657),\n",
       " ('relationship', 655),\n",
       " ('piece', 653),\n",
       " ('fantastic', 653),\n",
       " ('david', 651),\n",
       " ('turn', 648),\n",
       " ('murder', 646),\n",
       " ('parts', 645),\n",
       " ('brother', 644),\n",
       " ('head', 643),\n",
       " ('absolutely', 643),\n",
       " ('experience', 642),\n",
       " ('eyes', 641),\n",
       " ('sex', 638),\n",
       " ('called', 637),\n",
       " ('direction', 637),\n",
       " ('directed', 636),\n",
       " ('lines', 634),\n",
       " ('behind', 633),\n",
       " ('sort', 632),\n",
       " ('actress', 631),\n",
       " ('lead', 630),\n",
       " ('oscar', 628),\n",
       " ('example', 627),\n",
       " ('including', 627),\n",
       " ('known', 625),\n",
       " ('musical', 625),\n",
       " ('chance', 621),\n",
       " ('score', 620),\n",
       " ('feeling', 619),\n",
       " ('already', 619),\n",
       " ('hit', 619),\n",
       " ('voice', 615),\n",
       " ('living', 612),\n",
       " ('moment', 612),\n",
       " ('low', 610),\n",
       " ('supporting', 610),\n",
       " ('ago', 609),\n",
       " ('themselves', 608),\n",
       " ('reality', 605),\n",
       " ('hilarious', 605),\n",
       " ('jack', 604),\n",
       " ('told', 603),\n",
       " ('hand', 601),\n",
       " ('moving', 600),\n",
       " ('quality', 600),\n",
       " ('dialogue', 600),\n",
       " ('happy', 599),\n",
       " ('song', 599),\n",
       " ('matter', 598),\n",
       " ('paul', 598),\n",
       " ('light', 594),\n",
       " ('future', 593),\n",
       " ('entire', 592),\n",
       " ('finds', 591),\n",
       " ('gave', 589),\n",
       " ('laugh', 587),\n",
       " ('released', 586),\n",
       " ('expect', 584),\n",
       " ('fight', 581),\n",
       " ('particularly', 580),\n",
       " ('police', 579),\n",
       " ('cinematography', 579),\n",
       " ('whose', 578),\n",
       " ('sound', 578),\n",
       " ('type', 578),\n",
       " ('enjoyable', 573),\n",
       " ('view', 573),\n",
       " ('number', 572),\n",
       " ('husband', 572),\n",
       " ('romantic', 572),\n",
       " ('daughter', 572),\n",
       " ('documentary', 571),\n",
       " ('self', 570),\n",
       " ('robert', 569),\n",
       " ('superb', 569),\n",
       " ('took', 569),\n",
       " ('modern', 569),\n",
       " ('mean', 566),\n",
       " ('shown', 563),\n",
       " ('coming', 561),\n",
       " ('important', 560),\n",
       " ('king', 559),\n",
       " ('leave', 559),\n",
       " ('change', 558),\n",
       " ('somewhat', 555),\n",
       " ('wanted', 555),\n",
       " ('tells', 554),\n",
       " ('career', 552),\n",
       " ('country', 552),\n",
       " ('run', 552),\n",
       " ('events', 552),\n",
       " ('heard', 550),\n",
       " ('season', 550),\n",
       " ('girls', 549),\n",
       " ('greatest', 549),\n",
       " ('etc', 547),\n",
       " ('care', 546),\n",
       " ('starts', 545),\n",
       " ('english', 542),\n",
       " ('killer', 541),\n",
       " ('totally', 540),\n",
       " ('animation', 540),\n",
       " ('guys', 540),\n",
       " ('tale', 540),\n",
       " ('usual', 539),\n",
       " ('opinion', 535),\n",
       " ('miss', 535),\n",
       " ('easy', 531),\n",
       " ('violence', 531),\n",
       " ('songs', 530),\n",
       " ('british', 528),\n",
       " ('says', 526),\n",
       " ('realistic', 525),\n",
       " ('writing', 524),\n",
       " ('writer', 522),\n",
       " ('act', 522),\n",
       " ('comic', 521),\n",
       " ('thriller', 519),\n",
       " ('television', 517),\n",
       " ('power', 516),\n",
       " ('ones', 515),\n",
       " ('kid', 514),\n",
       " ('york', 513),\n",
       " ('novel', 513),\n",
       " ('problem', 512),\n",
       " ('alone', 512),\n",
       " ('attention', 509),\n",
       " ('involved', 508),\n",
       " ('kill', 507),\n",
       " ('extremely', 507),\n",
       " ('seemed', 506),\n",
       " ('hero', 505),\n",
       " ('french', 505),\n",
       " ('rock', 504),\n",
       " ('stuff', 501),\n",
       " ('wish', 499),\n",
       " ('begins', 498),\n",
       " ('taken', 497),\n",
       " ('sad', 497),\n",
       " ('ways', 496),\n",
       " ('richard', 495),\n",
       " ('knows', 494),\n",
       " ('atmosphere', 493),\n",
       " ('similar', 491),\n",
       " ('surprised', 491),\n",
       " ('car', 491),\n",
       " ('taking', 491),\n",
       " ('george', 490),\n",
       " ('perfectly', 490),\n",
       " ('across', 489),\n",
       " ('team', 489),\n",
       " ('sequence', 489),\n",
       " ('eye', 489),\n",
       " ('among', 488),\n",
       " ('powerful', 488),\n",
       " ('serious', 488),\n",
       " ('due', 488),\n",
       " ('room', 488),\n",
       " ('cannot', 487),\n",
       " ('order', 487),\n",
       " ('strange', 487),\n",
       " ('b', 487),\n",
       " ('beauty', 486),\n",
       " ('famous', 485),\n",
       " ('myself', 484),\n",
       " ('herself', 484),\n",
       " ('tries', 484),\n",
       " ('happened', 484),\n",
       " ('class', 483),\n",
       " ('four', 482),\n",
       " ('cool', 481),\n",
       " ('release', 479),\n",
       " ('theme', 479),\n",
       " ('anyway', 479),\n",
       " ('opening', 478),\n",
       " ('entertainment', 477),\n",
       " ('slow', 475),\n",
       " ('unique', 475),\n",
       " ('exactly', 475),\n",
       " ('ends', 475),\n",
       " ('red', 474),\n",
       " ('o', 474),\n",
       " ('level', 474),\n",
       " ('easily', 474),\n",
       " ('interest', 472),\n",
       " ('happen', 471),\n",
       " ('crime', 470),\n",
       " ('viewing', 468),\n",
       " ('sets', 467),\n",
       " ('memorable', 467),\n",
       " ('stop', 466),\n",
       " ('group', 466),\n",
       " ('working', 463),\n",
       " ('dance', 463),\n",
       " ('sister', 463),\n",
       " ('problems', 463),\n",
       " ('message', 463),\n",
       " ('knew', 462),\n",
       " ('mystery', 461),\n",
       " ('nature', 461),\n",
       " ('bring', 460),\n",
       " ('thinking', 459),\n",
       " ('brought', 459),\n",
       " ('believable', 459),\n",
       " ('mostly', 458),\n",
       " ('disney', 457),\n",
       " ('couldn', 457),\n",
       " ('society', 456),\n",
       " ('within', 455),\n",
       " ('lady', 455),\n",
       " ('blood', 454),\n",
       " ('parents', 453),\n",
       " ('upon', 453),\n",
       " ('viewers', 453),\n",
       " ('tom', 452),\n",
       " ('peter', 452),\n",
       " ('meets', 452),\n",
       " ('soundtrack', 452),\n",
       " ('form', 452),\n",
       " ('usually', 452),\n",
       " ('local', 450),\n",
       " ('follow', 448),\n",
       " ('certain', 448),\n",
       " ('whether', 447),\n",
       " ('possible', 446),\n",
       " ('emotional', 445),\n",
       " ('de', 444),\n",
       " ('above', 444),\n",
       " ('killed', 444),\n",
       " ('god', 443),\n",
       " ('middle', 443),\n",
       " ('happens', 442),\n",
       " ('needs', 442),\n",
       " ('flick', 442),\n",
       " ('masterpiece', 441),\n",
       " ('period', 440),\n",
       " ('major', 440),\n",
       " ('named', 439),\n",
       " ('haven', 439),\n",
       " ('particular', 438),\n",
       " ('th', 438),\n",
       " ('earth', 437),\n",
       " ('feature', 437),\n",
       " ('stand', 436),\n",
       " ('words', 435),\n",
       " ('typical', 435),\n",
       " ('obviously', 433),\n",
       " ('elements', 433),\n",
       " ('romance', 431),\n",
       " ('jane', 430),\n",
       " ('yourself', 427),\n",
       " ('showing', 427),\n",
       " ('fantasy', 426),\n",
       " ('brings', 426),\n",
       " ('guess', 423),\n",
       " ('america', 423),\n",
       " ('unfortunately', 422),\n",
       " ('huge', 422),\n",
       " ('indeed', 421),\n",
       " ('running', 421),\n",
       " ('talent', 420),\n",
       " ('stage', 419),\n",
       " ('started', 418),\n",
       " ('japanese', 417),\n",
       " ('leads', 417),\n",
       " ('sweet', 417),\n",
       " ('poor', 416),\n",
       " ('deal', 416),\n",
       " ('personal', 413),\n",
       " ('incredible', 413),\n",
       " ('fast', 412),\n",
       " ('deep', 410),\n",
       " ('became', 410),\n",
       " ('hours', 409),\n",
       " ('giving', 408),\n",
       " ('nearly', 408),\n",
       " ('dream', 408),\n",
       " ('turned', 407),\n",
       " ('clearly', 407),\n",
       " ('near', 406),\n",
       " ('obvious', 406),\n",
       " ('surprise', 405),\n",
       " ('cut', 405),\n",
       " ('body', 404),\n",
       " ('era', 404),\n",
       " ('five', 403),\n",
       " ('hour', 403),\n",
       " ('female', 403),\n",
       " ('note', 399),\n",
       " ('learn', 398),\n",
       " ('truth', 398),\n",
       " ('tony', 397),\n",
       " ('feels', 397),\n",
       " ('except', 397),\n",
       " ('match', 397),\n",
       " ('filmed', 394),\n",
       " ('clear', 394),\n",
       " ('complete', 394),\n",
       " ('older', 393),\n",
       " ('street', 393),\n",
       " ('eventually', 393),\n",
       " ('keeps', 393),\n",
       " ('lots', 393),\n",
       " ('buy', 392),\n",
       " ('stewart', 391),\n",
       " ('william', 391),\n",
       " ('joe', 390),\n",
       " ('meet', 390),\n",
       " ('fall', 390),\n",
       " ('rating', 389),\n",
       " ('unlike', 389),\n",
       " ('shots', 389),\n",
       " ('talking', 389),\n",
       " ('difficult', 389),\n",
       " ('means', 388),\n",
       " ('dramatic', 388),\n",
       " ('appears', 386),\n",
       " ('situation', 386),\n",
       " ('wonder', 386),\n",
       " ('subject', 386),\n",
       " ('present', 386),\n",
       " ('comments', 385),\n",
       " ('lee', 383),\n",
       " ('general', 383),\n",
       " ('sequences', 383),\n",
       " ('points', 382),\n",
       " ('earlier', 382),\n",
       " ('check', 379),\n",
       " ('gone', 379),\n",
       " ('suspense', 378),\n",
       " ('recommended', 378),\n",
       " ('ten', 378),\n",
       " ('business', 377),\n",
       " ('third', 377),\n",
       " ('leaves', 375),\n",
       " ('beyond', 375),\n",
       " ('talk', 375),\n",
       " ('portrayal', 374),\n",
       " ('beautifully', 373),\n",
       " ('single', 372),\n",
       " ('bill', 372),\n",
       " ('word', 371),\n",
       " ('plenty', 371),\n",
       " ('falls', 370),\n",
       " ('whom', 370),\n",
       " ('scary', 369),\n",
       " ('battle', 369),\n",
       " ('figure', 369),\n",
       " ('non', 369),\n",
       " ('return', 368),\n",
       " ('using', 368),\n",
       " ('doubt', 367),\n",
       " ('add', 367),\n",
       " ('hear', 366),\n",
       " ('solid', 366),\n",
       " ('success', 366),\n",
       " ('touching', 365),\n",
       " ('political', 365),\n",
       " ('oh', 365),\n",
       " ('jokes', 365),\n",
       " ('awesome', 364),\n",
       " ('hell', 364),\n",
       " ('boys', 364),\n",
       " ('recently', 362),\n",
       " ('dog', 362),\n",
       " ('sexual', 362),\n",
       " ('straight', 361),\n",
       " ('please', 361),\n",
       " ('wouldn', 361),\n",
       " ('features', 361),\n",
       " ('lack', 360),\n",
       " ('forget', 360),\n",
       " ('setting', 360),\n",
       " ('married', 359),\n",
       " ('mark', 359),\n",
       " ('social', 357),\n",
       " ('adventure', 356),\n",
       " ('interested', 356),\n",
       " ('terrific', 355),\n",
       " ('actual', 355),\n",
       " ('brothers', 355),\n",
       " ('sees', 355),\n",
       " ('call', 354),\n",
       " ('move', 354),\n",
       " ('various', 353),\n",
       " ('dr', 353),\n",
       " ('theater', 353),\n",
       " ('animated', 352),\n",
       " ('western', 351),\n",
       " ('baby', 350),\n",
       " ('space', 350),\n",
       " ('disappointed', 348),\n",
       " ('leading', 348),\n",
       " ('portrayed', 346),\n",
       " ('aren', 346),\n",
       " ('smith', 345),\n",
       " ('screenplay', 345),\n",
       " ('hate', 344),\n",
       " ('towards', 344),\n",
       " ('noir', 343),\n",
       " ('outstanding', 342),\n",
       " ('decent', 342),\n",
       " ('kelly', 342),\n",
       " ('journey', 341),\n",
       " ('directors', 341),\n",
       " ('effective', 340),\n",
       " ('looked', 340),\n",
       " ('none', 340),\n",
       " ('storyline', 339),\n",
       " ('mary', 339),\n",
       " ('cold', 339),\n",
       " ('fi', 339),\n",
       " ('caught', 339),\n",
       " ('sci', 339),\n",
       " ('rich', 338),\n",
       " ('charming', 338),\n",
       " ('manages', 337),\n",
       " ('rare', 337),\n",
       " ('popular', 337),\n",
       " ('harry', 337),\n",
       " ('spirit', 336),\n",
       " ('open', 335),\n",
       " ('appreciate', 335),\n",
       " ('basically', 334),\n",
       " ('moves', 334),\n",
       " ('acted', 334),\n",
       " ('pace', 333),\n",
       " ('subtle', 333),\n",
       " ('inside', 333),\n",
       " ('century', 333),\n",
       " ('deserves', 333),\n",
       " ('boring', 333),\n",
       " ('mention', 333),\n",
       " ('familiar', 332),\n",
       " ('background', 332),\n",
       " ('ben', 331),\n",
       " ('creepy', 330),\n",
       " ('supposed', 330),\n",
       " ('secret', 329),\n",
       " ('jim', 328),\n",
       " ('die', 328),\n",
       " ('question', 327),\n",
       " ('natural', 327),\n",
       " ('effect', 327),\n",
       " ('impressive', 326),\n",
       " ('language', 326),\n",
       " ('rate', 326),\n",
       " ('saying', 325),\n",
       " ('intelligent', 325),\n",
       " ('scott', 324),\n",
       " ('material', 324),\n",
       " ('telling', 324),\n",
       " ('realize', 324),\n",
       " ('singing', 323),\n",
       " ('dancing', 322),\n",
       " ('adult', 321),\n",
       " ('visual', 321),\n",
       " ('imagine', 321),\n",
       " ('kept', 320),\n",
       " ('office', 320),\n",
       " ('uses', 319),\n",
       " ('wait', 318),\n",
       " ('stunning', 318),\n",
       " ('pure', 318),\n",
       " ('previous', 317),\n",
       " ('copy', 317),\n",
       " ('seriously', 317),\n",
       " ('review', 317),\n",
       " ('hot', 316),\n",
       " ('reading', 316),\n",
       " ('created', 316),\n",
       " ('magic', 316),\n",
       " ('somehow', 316),\n",
       " ('create', 316),\n",
       " ('attempt', 315),\n",
       " ('air', 315),\n",
       " ('crazy', 315),\n",
       " ('stay', 315),\n",
       " ('escape', 315),\n",
       " ('frank', 315),\n",
       " ('hands', 314),\n",
       " ('filled', 313),\n",
       " ('surprisingly', 312),\n",
       " ('average', 312),\n",
       " ('expected', 312),\n",
       " ('complex', 311),\n",
       " ('studio', 310),\n",
       " ('quickly', 310),\n",
       " ('successful', 310),\n",
       " ('plus', 309),\n",
       " ('male', 309),\n",
       " ('co', 307),\n",
       " ('following', 306),\n",
       " ('images', 306),\n",
       " ('exciting', 306),\n",
       " ('minute', 306),\n",
       " ('casting', 306),\n",
       " ('themes', 305),\n",
       " ('e', 305),\n",
       " ('follows', 305),\n",
       " ('reasons', 305),\n",
       " ('members', 305),\n",
       " ('german', 305),\n",
       " ('cute', 304),\n",
       " ('genius', 304),\n",
       " ('free', 304),\n",
       " ('touch', 304),\n",
       " ('edge', 304),\n",
       " ('outside', 303),\n",
       " ('ok', 302),\n",
       " ('younger', 302),\n",
       " ('admit', 302),\n",
       " ('reviews', 302),\n",
       " ('master', 301),\n",
       " ('fighting', 301),\n",
       " ('odd', 301),\n",
       " ('thanks', 300),\n",
       " ('recent', 300),\n",
       " ('break', 300),\n",
       " ('comment', 300),\n",
       " ('apart', 299),\n",
       " ('lovely', 298),\n",
       " ('emotions', 298),\n",
       " ('begin', 298),\n",
       " ('doctor', 297),\n",
       " ('italian', 297),\n",
       " ('party', 297),\n",
       " ('missed', 296),\n",
       " ('south', 296),\n",
       " ...]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Examine the counts of the most common words in positive reviews\n",
    "positive_counts.most_common()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('', 561462),\n",
       " ('.', 167538),\n",
       " ('the', 163389),\n",
       " ('a', 79321),\n",
       " ('and', 74385),\n",
       " ('of', 69009),\n",
       " ('to', 68974),\n",
       " ('br', 52637),\n",
       " ('is', 50083),\n",
       " ('it', 48327),\n",
       " ('i', 46880),\n",
       " ('in', 43753),\n",
       " ('this', 40920),\n",
       " ('that', 37615),\n",
       " ('s', 31546),\n",
       " ('was', 26291),\n",
       " ('movie', 24965),\n",
       " ('for', 21927),\n",
       " ('but', 21781),\n",
       " ('with', 20878),\n",
       " ('as', 20625),\n",
       " ('t', 20361),\n",
       " ('film', 19218),\n",
       " ('you', 17549),\n",
       " ('on', 17192),\n",
       " ('not', 16354),\n",
       " ('have', 15144),\n",
       " ('are', 14623),\n",
       " ('be', 14541),\n",
       " ('he', 13856),\n",
       " ('one', 13134),\n",
       " ('they', 13011),\n",
       " ('at', 12279),\n",
       " ('his', 12147),\n",
       " ('all', 12036),\n",
       " ('so', 11463),\n",
       " ('like', 11238),\n",
       " ('there', 10775),\n",
       " ('just', 10619),\n",
       " ('by', 10549),\n",
       " ('or', 10272),\n",
       " ('an', 10266),\n",
       " ('who', 9969),\n",
       " ('from', 9731),\n",
       " ('if', 9518),\n",
       " ('about', 9061),\n",
       " ('out', 8979),\n",
       " ('what', 8422),\n",
       " ('some', 8306),\n",
       " ('no', 8143),\n",
       " ('her', 7947),\n",
       " ('even', 7687),\n",
       " ('can', 7653),\n",
       " ('has', 7604),\n",
       " ('good', 7423),\n",
       " ('bad', 7401),\n",
       " ('would', 7036),\n",
       " ('up', 6970),\n",
       " ('only', 6781),\n",
       " ('more', 6730),\n",
       " ('when', 6726),\n",
       " ('she', 6444),\n",
       " ('really', 6262),\n",
       " ('time', 6209),\n",
       " ('had', 6142),\n",
       " ('my', 6015),\n",
       " ('were', 6001),\n",
       " ('which', 5780),\n",
       " ('very', 5764),\n",
       " ('me', 5606),\n",
       " ('see', 5452),\n",
       " ('don', 5336),\n",
       " ('we', 5328),\n",
       " ('their', 5278),\n",
       " ('do', 5236),\n",
       " ('story', 5208),\n",
       " ('than', 5183),\n",
       " ('been', 5100),\n",
       " ('much', 5078),\n",
       " ('get', 5037),\n",
       " ('because', 4966),\n",
       " ('people', 4806),\n",
       " ('then', 4761),\n",
       " ('make', 4722),\n",
       " ('how', 4688),\n",
       " ('could', 4686),\n",
       " ('any', 4658),\n",
       " ('into', 4567),\n",
       " ('made', 4541),\n",
       " ('first', 4306),\n",
       " ('other', 4305),\n",
       " ('well', 4254),\n",
       " ('too', 4174),\n",
       " ('them', 4165),\n",
       " ('plot', 4154),\n",
       " ('movies', 4080),\n",
       " ('acting', 4056),\n",
       " ('will', 3993),\n",
       " ('way', 3989),\n",
       " ('most', 3919),\n",
       " ('him', 3858),\n",
       " ('after', 3838),\n",
       " ('its', 3655),\n",
       " ('think', 3643),\n",
       " ('also', 3608),\n",
       " ('characters', 3600),\n",
       " ('off', 3567),\n",
       " ('watch', 3550),\n",
       " ('character', 3506),\n",
       " ('did', 3506),\n",
       " ('why', 3463),\n",
       " ('being', 3393),\n",
       " ('better', 3358),\n",
       " ('know', 3334),\n",
       " ('over', 3316),\n",
       " ('seen', 3265),\n",
       " ('ever', 3263),\n",
       " ('never', 3259),\n",
       " ('your', 3233),\n",
       " ('where', 3219),\n",
       " ('two', 3173),\n",
       " ('little', 3096),\n",
       " ('films', 3077),\n",
       " ('here', 3027),\n",
       " ('m', 3000),\n",
       " ('nothing', 2990),\n",
       " ('say', 2982),\n",
       " ('end', 2954),\n",
       " ('something', 2942),\n",
       " ('should', 2920),\n",
       " ('many', 2909),\n",
       " ('does', 2871),\n",
       " ('thing', 2866),\n",
       " ('show', 2862),\n",
       " ('ve', 2829),\n",
       " ('scene', 2816),\n",
       " ('scenes', 2785),\n",
       " ('these', 2724),\n",
       " ('go', 2717),\n",
       " ('didn', 2646),\n",
       " ('great', 2640),\n",
       " ('watching', 2640),\n",
       " ('re', 2620),\n",
       " ('doesn', 2601),\n",
       " ('through', 2560),\n",
       " ('such', 2544),\n",
       " ('man', 2516),\n",
       " ('worst', 2480),\n",
       " ('actually', 2449),\n",
       " ('actors', 2437),\n",
       " ('life', 2429),\n",
       " ('back', 2424),\n",
       " ('while', 2418),\n",
       " ('director', 2405),\n",
       " ('funny', 2336),\n",
       " ('going', 2319),\n",
       " ('still', 2283),\n",
       " ('another', 2254),\n",
       " ('look', 2247),\n",
       " ('now', 2237),\n",
       " ('old', 2215),\n",
       " ('those', 2212),\n",
       " ('real', 2170),\n",
       " ('few', 2158),\n",
       " ('love', 2152),\n",
       " ('horror', 2150),\n",
       " ('before', 2147),\n",
       " ('want', 2141),\n",
       " ('minutes', 2126),\n",
       " ('pretty', 2115),\n",
       " ('best', 2094),\n",
       " ('though', 2091),\n",
       " ('same', 2081),\n",
       " ('script', 2074),\n",
       " ('work', 2027),\n",
       " ('every', 2025),\n",
       " ('seems', 2023),\n",
       " ('least', 2011),\n",
       " ('enough', 1997),\n",
       " ('down', 1988),\n",
       " ('original', 1983),\n",
       " ('guy', 1964),\n",
       " ('got', 1952),\n",
       " ('around', 1943),\n",
       " ('part', 1942),\n",
       " ('lot', 1892),\n",
       " ('anything', 1874),\n",
       " ('find', 1860),\n",
       " ('new', 1854),\n",
       " ('again', 1849),\n",
       " ('isn', 1849),\n",
       " ('point', 1845),\n",
       " ('fact', 1839),\n",
       " ('things', 1839),\n",
       " ('give', 1823),\n",
       " ('makes', 1814),\n",
       " ('take', 1800),\n",
       " ('thought', 1798),\n",
       " ('d', 1770),\n",
       " ('whole', 1768),\n",
       " ('long', 1761),\n",
       " ('years', 1759),\n",
       " ('however', 1740),\n",
       " ('gets', 1714),\n",
       " ('making', 1695),\n",
       " ('cast', 1694),\n",
       " ('big', 1662),\n",
       " ('might', 1658),\n",
       " ('interesting', 1648),\n",
       " ('money', 1638),\n",
       " ('us', 1628),\n",
       " ('right', 1625),\n",
       " ('far', 1619),\n",
       " ('quite', 1596),\n",
       " ('without', 1595),\n",
       " ('come', 1595),\n",
       " ('almost', 1574),\n",
       " ('ll', 1567),\n",
       " ('action', 1566),\n",
       " ('awful', 1557),\n",
       " ('kind', 1539),\n",
       " ('reason', 1534),\n",
       " ('am', 1530),\n",
       " ('looks', 1528),\n",
       " ('must', 1522),\n",
       " ('done', 1510),\n",
       " ('comedy', 1504),\n",
       " ('someone', 1490),\n",
       " ('trying', 1486),\n",
       " ('wasn', 1484),\n",
       " ('poor', 1481),\n",
       " ('instead', 1478),\n",
       " ('boring', 1478),\n",
       " ('saw', 1475),\n",
       " ('away', 1469),\n",
       " ('girl', 1463),\n",
       " ('probably', 1444),\n",
       " ('believe', 1434),\n",
       " ('sure', 1433),\n",
       " ('looking', 1430),\n",
       " ('stupid', 1428),\n",
       " ('anyone', 1418),\n",
       " ('times', 1406),\n",
       " ('world', 1404),\n",
       " ('maybe', 1404),\n",
       " ('rather', 1394),\n",
       " ('terrible', 1391),\n",
       " ('may', 1390),\n",
       " ('last', 1390),\n",
       " ('since', 1388),\n",
       " ('let', 1385),\n",
       " ('tv', 1382),\n",
       " ('hard', 1374),\n",
       " ('between', 1374),\n",
       " ('waste', 1358),\n",
       " ('woman', 1356),\n",
       " ('feel', 1354),\n",
       " ('effects', 1348),\n",
       " ('half', 1341),\n",
       " ('own', 1333),\n",
       " ('young', 1317),\n",
       " ('music', 1316),\n",
       " ('idea', 1312),\n",
       " ('sense', 1306),\n",
       " ('bit', 1298),\n",
       " ('having', 1280),\n",
       " ('book', 1278),\n",
       " ('found', 1267),\n",
       " ('put', 1263),\n",
       " ('series', 1263),\n",
       " ('goes', 1256),\n",
       " ('worse', 1249),\n",
       " ('said', 1230),\n",
       " ('comes', 1224),\n",
       " ('role', 1222),\n",
       " ('main', 1220),\n",
       " ('else', 1199),\n",
       " ('everything', 1197),\n",
       " ('yet', 1196),\n",
       " ('low', 1189),\n",
       " ('screen', 1188),\n",
       " ('supposed', 1186),\n",
       " ('actor', 1185),\n",
       " ('either', 1183),\n",
       " ('budget', 1179),\n",
       " ('ending', 1179),\n",
       " ('audience', 1178),\n",
       " ('set', 1177),\n",
       " ('family', 1170),\n",
       " ('left', 1169),\n",
       " ('completely', 1168),\n",
       " ('both', 1158),\n",
       " ('wrong', 1155),\n",
       " ('always', 1151),\n",
       " ('place', 1148),\n",
       " ('course', 1148),\n",
       " ('seem', 1147),\n",
       " ('watched', 1142),\n",
       " ('day', 1132),\n",
       " ('simply', 1130),\n",
       " ('shot', 1126),\n",
       " ('mean', 1117),\n",
       " ('special', 1102),\n",
       " ('dead', 1101),\n",
       " ('three', 1094),\n",
       " ('house', 1085),\n",
       " ('oh', 1084),\n",
       " ('night', 1083),\n",
       " ('read', 1082),\n",
       " ('less', 1067),\n",
       " ('high', 1066),\n",
       " ('year', 1064),\n",
       " ('camera', 1061),\n",
       " ('worth', 1057),\n",
       " ('our', 1056),\n",
       " ('try', 1051),\n",
       " ('sex', 1046),\n",
       " ('horrible', 1046),\n",
       " ('video', 1043),\n",
       " ('black', 1039),\n",
       " ('although', 1036),\n",
       " ('couldn', 1036),\n",
       " ('once', 1033),\n",
       " ('rest', 1022),\n",
       " ('dvd', 1021),\n",
       " ('line', 1018),\n",
       " ('played', 1017),\n",
       " ('fun', 1007),\n",
       " ('during', 1006),\n",
       " ('production', 1003),\n",
       " ('everyone', 1002),\n",
       " ('play', 993),\n",
       " ('mind', 990),\n",
       " ('kids', 989),\n",
       " ('version', 989),\n",
       " ('seeing', 988),\n",
       " ('american', 980),\n",
       " ('given', 978),\n",
       " ('used', 969),\n",
       " ('performance', 968),\n",
       " ('together', 963),\n",
       " ('especially', 963),\n",
       " ('tell', 959),\n",
       " ('women', 958),\n",
       " ('start', 956),\n",
       " ('need', 955),\n",
       " ('second', 953),\n",
       " ('takes', 950),\n",
       " ('each', 950),\n",
       " ('wife', 944),\n",
       " ('dialogue', 942),\n",
       " ('use', 940),\n",
       " ('problem', 938),\n",
       " ('star', 934),\n",
       " ('unfortunately', 931),\n",
       " ('himself', 929),\n",
       " ('doing', 926),\n",
       " ('death', 922),\n",
       " ('name', 921),\n",
       " ('lines', 919),\n",
       " ('killer', 914),\n",
       " ('getting', 913),\n",
       " ('help', 905),\n",
       " ('couple', 902),\n",
       " ('fan', 902),\n",
       " ('head', 898),\n",
       " ('crap', 895),\n",
       " ('guess', 888),\n",
       " ('piece', 884),\n",
       " ('nice', 880),\n",
       " ('different', 878),\n",
       " ('school', 876),\n",
       " ('later', 875),\n",
       " ('entire', 869),\n",
       " ('shows', 860),\n",
       " ('john', 858),\n",
       " ('next', 858),\n",
       " ('short', 857),\n",
       " ('seemed', 857),\n",
       " ('hollywood', 850),\n",
       " ('home', 848),\n",
       " ('true', 846),\n",
       " ('person', 846),\n",
       " ('absolutely', 842),\n",
       " ('sort', 840),\n",
       " ('care', 839),\n",
       " ('understand', 836),\n",
       " ('plays', 835),\n",
       " ('felt', 834),\n",
       " ('written', 829),\n",
       " ('title', 828),\n",
       " ('men', 822),\n",
       " ('until', 821),\n",
       " ('flick', 816),\n",
       " ('decent', 815),\n",
       " ('face', 814),\n",
       " ('friends', 810),\n",
       " ('job', 807),\n",
       " ('case', 807),\n",
       " ('stars', 807),\n",
       " ('itself', 804),\n",
       " ('yes', 801),\n",
       " ('perhaps', 800),\n",
       " ('went', 797),\n",
       " ('wanted', 797),\n",
       " ('called', 796),\n",
       " ('annoying', 795),\n",
       " ('tries', 790),\n",
       " ('ridiculous', 790),\n",
       " ('laugh', 788),\n",
       " ('evil', 787),\n",
       " ('along', 786),\n",
       " ('top', 785),\n",
       " ('hour', 784),\n",
       " ('full', 783),\n",
       " ('writing', 780),\n",
       " ('came', 780),\n",
       " ('keep', 770),\n",
       " ('totally', 767),\n",
       " ('playing', 766),\n",
       " ('god', 765),\n",
       " ('won', 764),\n",
       " ('guys', 763),\n",
       " ('already', 762),\n",
       " ('gore', 757),\n",
       " ('direction', 748),\n",
       " ('save', 746),\n",
       " ('lost', 745),\n",
       " ('example', 744),\n",
       " ('sound', 742),\n",
       " ('war', 741),\n",
       " ('attempt', 735),\n",
       " ('except', 733),\n",
       " ('car', 733),\n",
       " ('moments', 732),\n",
       " ('blood', 732),\n",
       " ('obviously', 730),\n",
       " ('act', 729),\n",
       " ('remember', 728),\n",
       " ('kill', 727),\n",
       " ('truly', 726),\n",
       " ('white', 726),\n",
       " ('father', 726),\n",
       " ('b', 725),\n",
       " ('thinking', 720),\n",
       " ('ok', 716),\n",
       " ('finally', 716),\n",
       " ('turn', 711),\n",
       " ('quality', 701),\n",
       " ('lack', 698),\n",
       " ('style', 694),\n",
       " ('wouldn', 693),\n",
       " ('cheap', 691),\n",
       " ('none', 690),\n",
       " ('please', 686),\n",
       " ('kid', 686),\n",
       " ('boy', 685),\n",
       " ('seriously', 684),\n",
       " ('lead', 680),\n",
       " ('dull', 677),\n",
       " ('children', 676),\n",
       " ('starts', 675),\n",
       " ('stuff', 673),\n",
       " ('hope', 672),\n",
       " ('looked', 670),\n",
       " ('recommend', 669),\n",
       " ('under', 668),\n",
       " ('run', 667),\n",
       " ('killed', 667),\n",
       " ('others', 666),\n",
       " ('enjoy', 666),\n",
       " ('etc', 663),\n",
       " ('myself', 663),\n",
       " ('against', 662),\n",
       " ('beginning', 662),\n",
       " ('girls', 662),\n",
       " ('small', 660),\n",
       " ('obvious', 660),\n",
       " ('hell', 659),\n",
       " ('slow', 657),\n",
       " ('hand', 656),\n",
       " ('wonder', 652),\n",
       " ('lame', 652),\n",
       " ('becomes', 651),\n",
       " ('picture', 651),\n",
       " ('based', 650),\n",
       " ('early', 648),\n",
       " ('behind', 646),\n",
       " ('poorly', 644),\n",
       " ('avoid', 642),\n",
       " ('apparently', 640),\n",
       " ('complete', 640),\n",
       " ('happens', 639),\n",
       " ('anyway', 638),\n",
       " ('classic', 637),\n",
       " ('several', 636),\n",
       " ('certainly', 635),\n",
       " ('despite', 635),\n",
       " ('episode', 635),\n",
       " ('often', 631),\n",
       " ('cut', 630),\n",
       " ('writer', 630),\n",
       " ('mother', 628),\n",
       " ('predictable', 628),\n",
       " ('gave', 628),\n",
       " ('become', 627),\n",
       " ('close', 625),\n",
       " ('fans', 624),\n",
       " ('saying', 621),\n",
       " ('scary', 619),\n",
       " ('stop', 618),\n",
       " ('live', 618),\n",
       " ('wants', 617),\n",
       " ('self', 615),\n",
       " ('mr', 612),\n",
       " ('friend', 611),\n",
       " ('jokes', 611),\n",
       " ('cannot', 610),\n",
       " ('overall', 609),\n",
       " ('cinema', 604),\n",
       " ('child', 603),\n",
       " ('silly', 601),\n",
       " ('beautiful', 596),\n",
       " ('human', 595),\n",
       " ('expect', 594),\n",
       " ('liked', 593),\n",
       " ('happened', 592),\n",
       " ('bunch', 590),\n",
       " ('entertaining', 590),\n",
       " ('final', 588),\n",
       " ('actress', 588),\n",
       " ('says', 584),\n",
       " ('performances', 584),\n",
       " ('turns', 577),\n",
       " ('humor', 577),\n",
       " ('eyes', 576),\n",
       " ('themselves', 576),\n",
       " ('hours', 574),\n",
       " ('happen', 573),\n",
       " ('days', 572),\n",
       " ('basically', 572),\n",
       " ('running', 571),\n",
       " ('disappointed', 569),\n",
       " ('involved', 569),\n",
       " ('call', 569),\n",
       " ('group', 568),\n",
       " ('directed', 568),\n",
       " ('fight', 567),\n",
       " ('body', 566),\n",
       " ('talking', 566),\n",
       " ('daughter', 566),\n",
       " ('badly', 565),\n",
       " ('sorry', 565),\n",
       " ('throughout', 563),\n",
       " ('viewer', 563),\n",
       " ('extremely', 562),\n",
       " ('yourself', 562),\n",
       " ('interest', 561),\n",
       " ('violence', 561),\n",
       " ('heard', 561),\n",
       " ('shots', 559),\n",
       " ('side', 557),\n",
       " ('word', 556),\n",
       " ('art', 555),\n",
       " ('possible', 554),\n",
       " ('dark', 551),\n",
       " ('game', 551),\n",
       " ('hero', 550),\n",
       " ('alone', 549),\n",
       " ('son', 547),\n",
       " ('leave', 547),\n",
       " ('type', 547),\n",
       " ('single', 546),\n",
       " ('gives', 546),\n",
       " ('parts', 546),\n",
       " ('started', 545),\n",
       " ('female', 543),\n",
       " ('voice', 541),\n",
       " ('mess', 541),\n",
       " ('rating', 541),\n",
       " ('town', 540),\n",
       " ('aren', 540),\n",
       " ('drama', 538),\n",
       " ('definitely', 537),\n",
       " ('unless', 536),\n",
       " ('review', 534),\n",
       " ('able', 533),\n",
       " ('weak', 533),\n",
       " ('effort', 533),\n",
       " ('took', 531),\n",
       " ('five', 530),\n",
       " ('non', 530),\n",
       " ('matter', 529),\n",
       " ('usually', 529),\n",
       " ('michael', 528),\n",
       " ('feeling', 526),\n",
       " ('huge', 523),\n",
       " ('sequel', 522),\n",
       " ('soon', 521),\n",
       " ('exactly', 520),\n",
       " ('past', 519),\n",
       " ('turned', 518),\n",
       " ('police', 518),\n",
       " ('tried', 515),\n",
       " ('middle', 513),\n",
       " ('talent', 513),\n",
       " ('genre', 512),\n",
       " ('zombie', 510),\n",
       " ('ends', 509),\n",
       " ('history', 509),\n",
       " ('straight', 503),\n",
       " ('coming', 501),\n",
       " ('serious', 501),\n",
       " ('opening', 501),\n",
       " ('moment', 500),\n",
       " ('sad', 499),\n",
       " ('lives', 499),\n",
       " ('particularly', 498),\n",
       " ('dialog', 498),\n",
       " ('editing', 493),\n",
       " ('clearly', 492),\n",
       " ('earth', 491),\n",
       " ('beyond', 491),\n",
       " ('taken', 490),\n",
       " ('cool', 490),\n",
       " ('level', 489),\n",
       " ('dumb', 489),\n",
       " ('okay', 488),\n",
       " ('major', 487),\n",
       " ('fast', 485),\n",
       " ('premise', 485),\n",
       " ('stories', 484),\n",
       " ('joke', 484),\n",
       " ('minute', 483),\n",
       " ('wasted', 483),\n",
       " ('mostly', 482),\n",
       " ('across', 482),\n",
       " ('rent', 482),\n",
       " ('late', 481),\n",
       " ('falls', 481),\n",
       " ('fails', 481),\n",
       " ('mention', 478),\n",
       " ('theater', 475),\n",
       " ('stay', 472),\n",
       " ('sometimes', 472),\n",
       " ('hit', 468),\n",
       " ('fine', 467),\n",
       " ('talk', 467),\n",
       " ('die', 466),\n",
       " ('storyline', 465),\n",
       " ('pointless', 465),\n",
       " ('taking', 464),\n",
       " ('order', 462),\n",
       " ('brother', 461),\n",
       " ('whatever', 460),\n",
       " ('told', 460),\n",
       " ('wish', 458),\n",
       " ('room', 456),\n",
       " ('appears', 455),\n",
       " ('write', 455),\n",
       " ('career', 455),\n",
       " ('husband', 454),\n",
       " ('known', 454),\n",
       " ('living', 451),\n",
       " ('ten', 450),\n",
       " ('sit', 450),\n",
       " ('words', 449),\n",
       " ('chance', 448),\n",
       " ('monster', 448),\n",
       " ('hate', 444),\n",
       " ('novel', 444),\n",
       " ('add', 443),\n",
       " ('english', 443),\n",
       " ('somehow', 441),\n",
       " ('strange', 440),\n",
       " ('imdb', 438),\n",
       " ('actual', 438),\n",
       " ('killing', 437),\n",
       " ('material', 437),\n",
       " ('total', 437),\n",
       " ('ones', 437),\n",
       " ('knew', 436),\n",
       " ('king', 434),\n",
       " ('number', 434),\n",
       " ('using', 433),\n",
       " ('lee', 431),\n",
       " ('shown', 431),\n",
       " ('works', 431),\n",
       " ('power', 431),\n",
       " ('giving', 431),\n",
       " ('possibly', 430),\n",
       " ('kept', 430),\n",
       " ('points', 430),\n",
       " ('four', 429),\n",
       " ('local', 427),\n",
       " ('usual', 426),\n",
       " ('including', 425),\n",
       " ('ago', 424),\n",
       " ('opinion', 424),\n",
       " ('problems', 424),\n",
       " ('nudity', 423),\n",
       " ('age', 422),\n",
       " ('due', 421),\n",
       " ('roles', 420),\n",
       " ('decided', 419),\n",
       " ('writers', 419),\n",
       " ('easily', 418),\n",
       " ('near', 418),\n",
       " ('flat', 418),\n",
       " ('experience', 417),\n",
       " ('murder', 417),\n",
       " ('reviews', 416),\n",
       " ('imagine', 415),\n",
       " ('feels', 413),\n",
       " ('somewhat', 411),\n",
       " ('plain', 411),\n",
       " ('score', 410),\n",
       " ('class', 410),\n",
       " ('song', 409),\n",
       " ('whether', 409),\n",
       " ('bring', 409),\n",
       " ('otherwise', 408),\n",
       " ('average', 408),\n",
       " ('whose', 408),\n",
       " ('pathetic', 407),\n",
       " ('nearly', 407),\n",
       " ('zombies', 407),\n",
       " ('knows', 407),\n",
       " ('cheesy', 406),\n",
       " ('upon', 406),\n",
       " ('cinematography', 406),\n",
       " ('city', 405),\n",
       " ('space', 405),\n",
       " ('credits', 404),\n",
       " ('change', 403),\n",
       " ('james', 403),\n",
       " ('lots', 403),\n",
       " ('entertainment', 402),\n",
       " ('nor', 402),\n",
       " ('wait', 401),\n",
       " ('released', 400),\n",
       " ('needs', 399),\n",
       " ('shame', 398),\n",
       " ('attention', 396),\n",
       " ('comments', 394),\n",
       " ('lady', 393),\n",
       " ('free', 393),\n",
       " ('bored', 393),\n",
       " ('clear', 392),\n",
       " ('expected', 392),\n",
       " ('needed', 392),\n",
       " ('view', 391),\n",
       " ('doubt', 390),\n",
       " ('check', 390),\n",
       " ('development', 390),\n",
       " ('mystery', 389),\n",
       " ('figure', 389),\n",
       " ('garbage', 388),\n",
       " ('excellent', 388),\n",
       " ('television', 386),\n",
       " ('sequence', 386),\n",
       " ('sets', 385),\n",
       " ('o', 385),\n",
       " ('potential', 384),\n",
       " ('laughable', 384),\n",
       " ('reality', 382),\n",
       " ('documentary', 382),\n",
       " ('robert', 382),\n",
       " ('light', 382),\n",
       " ('country', 382),\n",
       " ('general', 381),\n",
       " ('ask', 381),\n",
       " ('begin', 380),\n",
       " ('comic', 380),\n",
       " ('fall', 380),\n",
       " ('trash', 379),\n",
       " ('forced', 379),\n",
       " ('footage', 379),\n",
       " ('stand', 379),\n",
       " ('remake', 379),\n",
       " ('songs', 378),\n",
       " ('thriller', 378),\n",
       " ('within', 377),\n",
       " ('gay', 377),\n",
       " ('hardly', 376),\n",
       " ('gone', 375),\n",
       " ('above', 375),\n",
       " ('george', 374),\n",
       " ('sounds', 373),\n",
       " ('means', 373),\n",
       " ('buy', 372),\n",
       " ('david', 372),\n",
       " ('directing', 372),\n",
       " ('move', 372),\n",
       " ('important', 371),\n",
       " ('rock', 371),\n",
       " ('forward', 371),\n",
       " ('hot', 370),\n",
       " ('filmed', 370),\n",
       " ('british', 370),\n",
       " ('haven', 370),\n",
       " ('reading', 369),\n",
       " ('heart', 369),\n",
       " ('fake', 369),\n",
       " ('hear', 368),\n",
       " ('weird', 368),\n",
       " ('incredibly', 368),\n",
       " ('cop', 367),\n",
       " ('enjoyed', 367),\n",
       " ('hilarious', 367),\n",
       " ('musical', 367),\n",
       " ('happy', 366),\n",
       " ('pay', 366),\n",
       " ('message', 366),\n",
       " ('laughs', 365),\n",
       " ('box', 365),\n",
       " ('suspense', 363),\n",
       " ('sadly', 363),\n",
       " ('eye', 362),\n",
       " ('similar', 361),\n",
       " ('named', 361),\n",
       " ('third', 361),\n",
       " ('modern', 360),\n",
       " ('failed', 359),\n",
       " ('events', 359),\n",
       " ('question', 358),\n",
       " ('forget', 358),\n",
       " ('finds', 357),\n",
       " ('male', 357),\n",
       " ('perfect', 356),\n",
       " ('spent', 355),\n",
       " ('sister', 355),\n",
       " ('feature', 354),\n",
       " ('result', 354),\n",
       " ('girlfriend', 353),\n",
       " ('comment', 353),\n",
       " ('sexual', 352),\n",
       " ('richard', 351),\n",
       " ('neither', 351),\n",
       " ('attempts', 351),\n",
       " ('screenplay', 350),\n",
       " ('elements', 350),\n",
       " ('spoilers', 349),\n",
       " ('brain', 348),\n",
       " ('filmmakers', 348),\n",
       " ('showing', 348),\n",
       " ('dr', 347),\n",
       " ('christmas', 347),\n",
       " ('miss', 347),\n",
       " ('cover', 345),\n",
       " ('red', 344),\n",
       " ('sequences', 344),\n",
       " ('excuse', 343),\n",
       " ('typical', 343),\n",
       " ('baby', 342),\n",
       " ('ideas', 342),\n",
       " ('crazy', 342),\n",
       " ('loved', 341),\n",
       " ('meant', 341),\n",
       " ('fire', 340),\n",
       " ('worked', 340),\n",
       " ('follow', 339),\n",
       " ('unbelievable', 339),\n",
       " ('theme', 337),\n",
       " ('producers', 336),\n",
       " ('twist', 336),\n",
       " ('plus', 336),\n",
       " ('barely', 336),\n",
       " ('appear', 336),\n",
       " ('team', 335),\n",
       " ('directors', 335),\n",
       " ('viewers', 333),\n",
       " ('slasher', 332),\n",
       " ('tom', 332),\n",
       " ('leads', 332),\n",
       " ('wrote', 331),\n",
       " ('working', 331),\n",
       " ('villain', 331),\n",
       " ('gun', 331),\n",
       " ('open', 330),\n",
       " ('island', 330),\n",
       " ('strong', 330),\n",
       " ('realize', 330),\n",
       " ('yeah', 329),\n",
       " ('positive', 329),\n",
       " ('quickly', 329),\n",
       " ('disappointing', 329),\n",
       " ('weren', 328),\n",
       " ('simple', 328),\n",
       " ('honestly', 328),\n",
       " ('release', 328),\n",
       " ('period', 327),\n",
       " ('tells', 327),\n",
       " ('doctor', 327),\n",
       " ('eventually', 327),\n",
       " ('kills', 327),\n",
       " ('list', 326),\n",
       " ('nowhere', 326),\n",
       " ('dog', 326),\n",
       " ('herself', 326),\n",
       " ('acted', 326),\n",
       " ('walk', 325),\n",
       " ('apart', 324),\n",
       " ('air', 324),\n",
       " ('subject', 323),\n",
       " ('makers', 323),\n",
       " ('learn', 322),\n",
       " ('fi', 322),\n",
       " ('bother', 319),\n",
       " ('sci', 319),\n",
       " ('admit', 319),\n",
       " ('disappointment', 318),\n",
       " ('note', 318),\n",
       " ('jack', 318),\n",
       " ('hands', 318),\n",
       " ('e', 317),\n",
       " ('value', 317),\n",
       " ('casting', 317),\n",
       " ('certain', 317),\n",
       " ('grade', 316),\n",
       " ('peter', 316),\n",
       " ('missing', 315),\n",
       " ('suddenly', 315),\n",
       " ('stick', 313),\n",
       " ('previous', 313),\n",
       " ('form', 313),\n",
       " ('break', 313),\n",
       " ('soundtrack', 312),\n",
       " ('surprised', 311),\n",
       " ('front', 311),\n",
       " ('expecting', 311),\n",
       " ('parents', 310),\n",
       " ('surprise', 310),\n",
       " ('relationship', 310),\n",
       " ('shoot', 309),\n",
       " ('today', 309),\n",
       " ('ended', 308),\n",
       " ('creepy', 308),\n",
       " ('ways', 308),\n",
       " ('somewhere', 308),\n",
       " ('concept', 308),\n",
       " ('leaves', 308),\n",
       " ('vampire', 308),\n",
       " ('painful', 308),\n",
       " ('spend', 307),\n",
       " ('th', 307),\n",
       " ('future', 306),\n",
       " ('street', 306),\n",
       " ('fighting', 306),\n",
       " ('effect', 306),\n",
       " ('difficult', 306),\n",
       " ('america', 305),\n",
       " ('c', 305),\n",
       " ('accent', 304),\n",
       " ('project', 302),\n",
       " ('truth', 302),\n",
       " ('indeed', 301),\n",
       " ('joe', 301),\n",
       " ('deal', 301),\n",
       " ('f', 301),\n",
       " ('biggest', 300),\n",
       " ('rate', 300),\n",
       " ('paul', 299),\n",
       " ('japanese', 299),\n",
       " ('utterly', 298),\n",
       " ('college', 298),\n",
       " ('begins', 298),\n",
       " ('redeeming', 298),\n",
       " ('disney', 297),\n",
       " ('york', 297),\n",
       " ('fairly', 297),\n",
       " ('crew', 296),\n",
       " ('revenge', 296),\n",
       " ('create', 296),\n",
       " ('cartoon', 296),\n",
       " ('outside', 295),\n",
       " ('co', 295),\n",
       " ('computer', 295),\n",
       " ('stage', 295),\n",
       " ('interested', 295),\n",
       " ('among', 294),\n",
       " ('speak', 294),\n",
       " ('considering', 294),\n",
       " ('channel', 293),\n",
       " ('sick', 293),\n",
       " ('towards', 293),\n",
       " ('talented', 292),\n",
       " ('particular', 292),\n",
       " ('hair', 292),\n",
       " ('van', 292),\n",
       " ('cause', 292),\n",
       " ('bottom', 291),\n",
       " ('reasons', 291),\n",
       " ('cat', 290),\n",
       " ('mediocre', 290),\n",
       " ('telling', 290),\n",
       " ('store', 289),\n",
       " ('supporting', 289),\n",
       " ('hoping', 288),\n",
       " ('waiting', 288),\n",
       " ...]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Examine the counts of the most common words in negative reviews\n",
    "negative_counts.most_common()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, common words like \"the\" appear very often in both positive and negative reviews. Instead of finding the most common words in positive or negative reviews, what you really want are the words found in positive reviews more often than in negative reviews, and vice versa. To accomplish this, you'll need to calculate the **ratios** of word usage between positive and negative reviews.\n",
    "\n",
    "**TODO:** Check all the words you've seen and calculate the ratio of postive to negative uses and store that ratio in `pos_neg_ratios`. \n",
    ">Hint: the positive-to-negative ratio for a given word can be calculated with `positive_counts[word] / float(negative_counts[word]+1)`. Notice the `+1` in the denominator – that ensures we don't divide by zero for words that are only seen in positive reviews."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Create Counter object to store positive/negative ratios\n",
    "pos_neg_ratios = Counter()\n",
    "\n",
    "# TODO: Calculate the ratios of positive and negative uses of the most common words\n",
    "#       Consider words to be \"common\" if they've been used at least 100 times\n",
    "for word in total_counts:\n",
    "    pos_neg_ratios[word] = positive_counts[word] / float(negative_counts[word]+1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Examine the ratios you've calculated for a few words:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pos-to-neg ratio for 'the' = 1.0607993145235326\n",
      "Pos-to-neg ratio for 'amazing' = 4.022813688212928\n",
      "Pos-to-neg ratio for 'terrible' = 0.17744252873563218\n"
     ]
    }
   ],
   "source": [
    "print(\"Pos-to-neg ratio for 'the' = {}\".format(pos_neg_ratios[\"the\"]))\n",
    "print(\"Pos-to-neg ratio for 'amazing' = {}\".format(pos_neg_ratios[\"amazing\"]))\n",
    "print(\"Pos-to-neg ratio for 'terrible' = {}\".format(pos_neg_ratios[\"terrible\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking closely at the values you just calculated, we see the following:\n",
    "\n",
    "* Words that you would expect to see more often in positive reviews – like \"amazing\" – have a ratio greater than 1. The more skewed a word is toward postive, the farther from 1 its positive-to-negative ratio  will be.\n",
    "* Words that you would expect to see more often in negative reviews – like \"terrible\" – have positive values that are less than 1. The more skewed a word is toward negative, the closer to zero its positive-to-negative ratio will be.\n",
    "* Neutral words, which don't really convey any sentiment because you would expect to see them in all sorts of reviews – like \"the\" – have values very close to 1. A perfectly neutral word – one that was used in exactly the same number of positive reviews as negative reviews – would be almost exactly 1. The `+1` we suggested you add to the denominator slightly biases words toward negative, but it won't matter because it will be a tiny bias and later we'll be ignoring words that are too close to neutral anyway.\n",
    "\n",
    "Ok, the ratios tell us which words are used more often in postive or negative reviews, but the specific values we've calculated are a bit difficult to work with. A very positive word like \"amazing\" has a value above 4, whereas a very negative word like \"terrible\" has a value around 0.18. Those values aren't easy to compare for a couple of reasons:\n",
    "\n",
    "* Right now, 1 is considered neutral, but the absolute value of the postive-to-negative rations of very postive words is larger than the absolute value of the ratios for the very negative words. So there is no way to directly compare two numbers and see if one word conveys the same magnitude of positive sentiment as another word conveys negative sentiment. So we should center all the values around netural so the absolute value fro neutral of the postive-to-negative ratio for a word would indicate how much sentiment (positive or negative) that word conveys.\n",
    "* When comparing absolute values it's easier to do that around zero than one. \n",
    "\n",
    "To fix these issues, we'll convert all of our ratios to new values using logarithms.\n",
    "\n",
    "**TODO:** Go through all the ratios you calculated and convert their values using the following formulas:\n",
    "> * For any postive words, convert the ratio using `np.log(ratio)`\n",
    "> * For any negative words, convert the ratio using `-np.log(1/(ratio + 0.01))`\n",
    "\n",
    "That second equation may look strange, but what it's doing is dividing one by a very small number, which will produce a larger positive number. Then, it takes the `log` of that, which produces numbers similar to the ones for the postive words. Finally, we negate the values by adding that minus sign up front. In the end, extremely positive and extremely negative words will have positive-to-negative ratios with similar magnitudes but oppositite signs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# TODO: Convert ratios to logs\n",
    "for word,ratio in pos_neg_ratios.most_common():\n",
    "    if (ratio > 1.0):\n",
    "        pos_neg_ratios[word] = np.log(ratio)\n",
    "    else:\n",
    "        pos_neg_ratios[word] = -np.log(1/(ratio+0.01))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Examine the new ratios you've calculated for the same words from before:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pos-to-neg ratio for 'the' = 0.05902269426102881\n",
      "Pos-to-neg ratio for 'amazing' = 1.3919815802404802\n",
      "Pos-to-neg ratio for 'terrible' = -1.6742829939664696\n"
     ]
    }
   ],
   "source": [
    "print(\"Pos-to-neg ratio for 'the' = {}\".format(pos_neg_ratios[\"the\"]))\n",
    "print(\"Pos-to-neg ratio for 'amazing' = {}\".format(pos_neg_ratios[\"amazing\"]))\n",
    "print(\"Pos-to-neg ratio for 'terrible' = {}\".format(pos_neg_ratios[\"terrible\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If everything worked, now you should see neutral words with values close to zero. In this case, \"the\" is near zero but slightly positive, so it was probably used in more positive reviews than negative reviews. But look at \"amazing\"'s ratio - it's above `1`, showing it is clearly a word with positive sentiment. And \"terrible\" has a similar score, but in the opposite direction, so it's below `-1`. It's now clear that both of these words are associated with specific, opposing sentiments.\n",
    "\n",
    "Now run the following cells to see more ratios. \n",
    "\n",
    "The first cell displays all the words, ordered by how associated they are with postive reviews. (Your notebook will most likely truncate the output so you won't actually see *all* the words in the list.)\n",
    "\n",
    "The second cell displays the 30 words most associated with negative reviews by reversing the order of the first list and then looking at the first 30 words. (If you want the second cell to display all the words, ordered by how associated they are with negative reviews, you could just write `reversed(pos_neg_ratios.most_common())`.)\n",
    "\n",
    "You should continue to see values similar to the earlier ones we checked – neutral words will be close to `0`, words will get more positive as their ratios approach and go above `1`, and words will get more negative as their ratios approach and go below `-1`. That's why we decided to use the logs instead of the raw ratios."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('edie', 4.6913478822291435),\n",
       " ('antwone', 4.4773368144782069),\n",
       " ('din', 4.4067192472642533),\n",
       " ('gunga', 4.1896547420264252),\n",
       " ('goldsworthy', 4.1743872698956368),\n",
       " ('yokai', 4.0943445622221004),\n",
       " ('gypo', 4.0943445622221004),\n",
       " ('paulie', 4.0775374439057197),\n",
       " ('visconti', 3.9318256327243257),\n",
       " ('flavia', 3.9318256327243257),\n",
       " ('kells', 3.8712010109078911),\n",
       " ('blandings', 3.8712010109078911),\n",
       " ('brashear', 3.8501476017100584),\n",
       " ('gino', 3.8286413964890951),\n",
       " ('deathtrap', 3.8066624897703196),\n",
       " ('panahi', 3.713572066704308),\n",
       " ('harilal', 3.713572066704308),\n",
       " ('ossessione', 3.6635616461296463),\n",
       " ('caruso', 3.6375861597263857),\n",
       " ('tsui', 3.6375861597263857),\n",
       " ('ahmad', 3.6109179126442243),\n",
       " ('sabu', 3.6109179126442243),\n",
       " ('khouri', 3.5835189384561099),\n",
       " ('dominick', 3.5835189384561099),\n",
       " ('aweigh', 3.5553480614894135),\n",
       " ('mj', 3.5553480614894135),\n",
       " ('kriemhild', 3.5263605246161616),\n",
       " ('mcintire', 3.5263605246161616),\n",
       " ('daisies', 3.4965075614664802),\n",
       " ('newcombe', 3.4965075614664802),\n",
       " ('blackie', 3.4965075614664802),\n",
       " ('trelkovsky', 3.4657359027997265),\n",
       " ('kei', 3.4657359027997265),\n",
       " ('jaffar', 3.4339872044851463),\n",
       " ('hilliard', 3.4339872044851463),\n",
       " ('gundam', 3.4231762883809305),\n",
       " ('offside', 3.4011973816621555),\n",
       " ('pazu', 3.4011973816621555),\n",
       " ('bathsheba', 3.4011973816621555),\n",
       " ('sheeta', 3.4011973816621555),\n",
       " ('krell', 3.4011973816621555),\n",
       " ('venoms', 3.3672958299864741),\n",
       " ('fineman', 3.3672958299864741),\n",
       " ('paine', 3.3322045101752038),\n",
       " ('ronny', 3.3322045101752038),\n",
       " ('ranma', 3.3322045101752038),\n",
       " ('pimlico', 3.3322045101752038),\n",
       " ('abhay', 3.2958368660043291),\n",
       " ('iturbi', 3.2771447329921766),\n",
       " ('kipling', 3.2580965380214821),\n",
       " ('pym', 3.2580965380214821),\n",
       " ('audiard', 3.2188758248682006),\n",
       " ('kelso', 3.2188758248682006),\n",
       " ('gabe', 3.2188758248682006),\n",
       " ('milverton', 3.2188758248682006),\n",
       " ('scalise', 3.2188758248682006),\n",
       " ('giovanna', 3.1780538303479458),\n",
       " ('moonwalker', 3.1780538303479458),\n",
       " ('grisby', 3.1780538303479458),\n",
       " ('xica', 3.1780538303479458),\n",
       " ('mukhsin', 3.1780538303479458),\n",
       " ('feinstone', 3.1780538303479458),\n",
       " ('felix', 3.1527360223636558),\n",
       " ('jannings', 3.1354942159291497),\n",
       " ('togar', 3.1354942159291497),\n",
       " ('heaton', 3.1354942159291497),\n",
       " ('chikatilo', 3.1354942159291497),\n",
       " ('luzhin', 3.1135153092103742),\n",
       " ('desdemona', 3.0910424533583161),\n",
       " ('miklos', 3.0910424533583161),\n",
       " ('matuschek', 3.0910424533583161),\n",
       " ('leonora', 3.0910424533583161),\n",
       " ('soha', 3.0910424533583161),\n",
       " ('fanfan', 3.0910424533583161),\n",
       " ('pidgeon', 3.0910424533583161),\n",
       " ('matador', 3.0910424533583161),\n",
       " ('philo', 3.068052935133617),\n",
       " ('lindy', 3.044522437723423),\n",
       " ('bjm', 3.044522437723423),\n",
       " ('gauri', 3.044522437723423),\n",
       " ('microfilm', 3.044522437723423),\n",
       " ('reda', 3.044522437723423),\n",
       " ('firemen', 3.044522437723423),\n",
       " ('maradona', 3.044522437723423),\n",
       " ('joss', 3.044522437723423),\n",
       " ('capote', 3.0122615755052013),\n",
       " ('fido', 3.0081547935525483),\n",
       " ('emory', 2.9957322735539909),\n",
       " ('carrre', 2.9957322735539909),\n",
       " ('quibble', 2.9957322735539909),\n",
       " ('siegfried', 2.9957322735539909),\n",
       " ('coonskin', 2.9957322735539909),\n",
       " ('mcintyre', 2.9957322735539909),\n",
       " ('prote', 2.9957322735539909),\n",
       " ('coe', 2.9957322735539909),\n",
       " ('excellently', 2.9789251552376097),\n",
       " ('clutter', 2.9704144655697009),\n",
       " ('anchors', 2.9444389791664403),\n",
       " ('railly', 2.9444389791664403),\n",
       " ('malfique', 2.9444389791664403),\n",
       " ('hillyer', 2.9444389791664403),\n",
       " ('pinjar', 2.9444389791664403),\n",
       " ('pakeezah', 2.9444389791664403),\n",
       " ('ferdie', 2.9444389791664403),\n",
       " ('schlesinger', 2.9444389791664403),\n",
       " ('vance', 2.9444389791664403),\n",
       " ('baloo', 2.9444389791664403),\n",
       " ('petiot', 2.9444389791664403),\n",
       " ('versatility', 2.9444389791664403),\n",
       " ('knockout', 2.9444389791664403),\n",
       " ('rvd', 2.9444389791664403),\n",
       " ('magnus', 2.9444389791664403),\n",
       " ('ackland', 2.9444389791664403),\n",
       " ('burakov', 2.9444389791664403),\n",
       " ('flippen', 2.9444389791664403),\n",
       " ('digicorp', 2.9444389791664403),\n",
       " ('kolchak', 2.9311937524164198),\n",
       " ('deanna', 2.9267394020670396),\n",
       " ('hayworth', 2.9267394020670396),\n",
       " ('hobson', 2.8903717578961645),\n",
       " ('hoechlin', 2.8903717578961645),\n",
       " ('pollak', 2.8903717578961645),\n",
       " ('geer', 2.8903717578961645),\n",
       " ('iphigenia', 2.8903717578961645),\n",
       " ('lando', 2.8903717578961645),\n",
       " ('pappas', 2.8903717578961645),\n",
       " ('guerrero', 2.8903717578961645),\n",
       " ('falco', 2.8903717578961645),\n",
       " ('burgade', 2.8903717578961645),\n",
       " ('volckman', 2.8903717578961645),\n",
       " ('iek', 2.8903717578961645),\n",
       " ('korda', 2.8622008809294686),\n",
       " ('sammo', 2.8526314299133175),\n",
       " ('orked', 2.8332133440562162),\n",
       " ('nighy', 2.8332133440562162),\n",
       " ('cdric', 2.8332133440562162),\n",
       " ('biko', 2.8332133440562162),\n",
       " ('hewlett', 2.8332133440562162),\n",
       " ('nagra', 2.8332133440562162),\n",
       " ('callahan', 2.8332133440562162),\n",
       " ('naudet', 2.8332133440562162),\n",
       " ('kralik', 2.8332133440562162),\n",
       " ('peralta', 2.8332133440562162),\n",
       " ('jeon', 2.8332133440562162),\n",
       " ('beckett', 2.8332133440562162),\n",
       " ('jacknife', 2.8332133440562162),\n",
       " ('sox', 2.8332133440562162),\n",
       " ('toughness', 2.8332133440562162),\n",
       " ('laputa', 2.8332133440562162),\n",
       " ('polanski', 2.8233610476132043),\n",
       " ('alvin', 2.8183982582710754),\n",
       " ('matthau', 2.8067217286092401),\n",
       " ('aiello', 2.8033603809065348),\n",
       " ('hulce', 2.7725887222397811),\n",
       " ('antz', 2.7725887222397811),\n",
       " ('gaiman', 2.7725887222397811),\n",
       " ('endor', 2.7725887222397811),\n",
       " ('mathieu', 2.7725887222397811),\n",
       " ('delpy', 2.7725887222397811),\n",
       " ('santos', 2.7725887222397811),\n",
       " ('janos', 2.7725887222397811),\n",
       " ('firefighter', 2.7725887222397811),\n",
       " ('macready', 2.7725887222397811),\n",
       " ('rotj', 2.7725887222397811),\n",
       " ('yvaine', 2.7725887222397811),\n",
       " ('duffell', 2.7725887222397811),\n",
       " ('bedknobs', 2.7725887222397811),\n",
       " ('bombshells', 2.7725887222397811),\n",
       " ('coulouris', 2.7725887222397811),\n",
       " ('natalia', 2.7725887222397811),\n",
       " ('myrtle', 2.7515353130419489),\n",
       " ('adele', 2.7515353130419489),\n",
       " ('gackt', 2.7515353130419489),\n",
       " ('bake', 2.7408400239252009),\n",
       " ('gilliam', 2.7245795030534206),\n",
       " ('valette', 2.7080502011022101),\n",
       " ('goines', 2.7080502011022101),\n",
       " ('mcanally', 2.7080502011022101),\n",
       " ('grasshoppers', 2.7080502011022101),\n",
       " ('lian', 2.7080502011022101),\n",
       " ('schildkraut', 2.7080502011022101),\n",
       " ('aviv', 2.7080502011022101),\n",
       " ('oro', 2.7080502011022101),\n",
       " ('bernsen', 2.7080502011022101),\n",
       " ('shintaro', 2.7080502011022101),\n",
       " ('hilda', 2.7080502011022101),\n",
       " ('cheh', 2.7080502011022101),\n",
       " ('baton', 2.7080502011022101),\n",
       " ('johnnie', 2.7080502011022101),\n",
       " ('soutendijk', 2.7080502011022101),\n",
       " ('lanisha', 2.7080502011022101),\n",
       " ('ashraf', 2.7080502011022101),\n",
       " ('parador', 2.7080502011022101),\n",
       " ('hickock', 2.7080502011022101),\n",
       " ('karas', 2.7080502011022101),\n",
       " ('eustache', 2.7080502011022101),\n",
       " ('kiley', 2.7080502011022101),\n",
       " ('doktor', 2.7080502011022101),\n",
       " ('burman', 2.7080502011022101),\n",
       " ('calamai', 2.7080502011022101),\n",
       " ('silberling', 2.7080502011022101),\n",
       " ('gunbuster', 2.7080502011022101),\n",
       " ('cognac', 2.7080502011022101),\n",
       " ('ballantine', 2.7080502011022101),\n",
       " ('uld', 2.7080502011022101),\n",
       " ('gannon', 2.7080502011022101),\n",
       " ('rotoscoped', 2.7080502011022101),\n",
       " ('egon', 2.7080502011022101),\n",
       " ('victoria', 2.6810215287142909),\n",
       " ('dench', 2.6741486494265287),\n",
       " ('partition', 2.6741486494265287),\n",
       " ('beery', 2.6741486494265287),\n",
       " ('conroy', 2.6741486494265287),\n",
       " ('chavez', 2.6672282065819548),\n",
       " ('ratso', 2.653241964607215),\n",
       " ('beek', 2.6390573296152584),\n",
       " ('melancholic', 2.6390573296152584),\n",
       " ('ishwar', 2.6390573296152584),\n",
       " ('fetisov', 2.6390573296152584),\n",
       " ('intricately', 2.6390573296152584),\n",
       " ('megs', 2.6390573296152584),\n",
       " ('bathhouse', 2.6390573296152584),\n",
       " ('hickam', 2.6390573296152584),\n",
       " ('kulkarni', 2.6390573296152584),\n",
       " ('tulip', 2.6390573296152584),\n",
       " ('scrat', 2.6390573296152584),\n",
       " ('pilgrimage', 2.6390573296152584),\n",
       " ('gialli', 2.6390573296152584),\n",
       " ('atoz', 2.6390573296152584),\n",
       " ('ingram', 2.6390573296152584),\n",
       " ('duprez', 2.6390573296152584),\n",
       " ('rien', 2.6390573296152584),\n",
       " ('harriet', 2.6390573296152584),\n",
       " ('danelia', 2.6390573296152584),\n",
       " ('emy', 2.6390573296152584),\n",
       " ('cartwrights', 2.6390573296152584),\n",
       " ('pang', 2.6390573296152584),\n",
       " ('gardenia', 2.6390573296152584),\n",
       " ('laine', 2.6390573296152584),\n",
       " ('treaty', 2.6390573296152584),\n",
       " ('cb', 2.6390573296152584),\n",
       " ('kabei', 2.6390573296152584),\n",
       " ('embezzler', 2.6390573296152584),\n",
       " ('mildred', 2.6026896854443837),\n",
       " ('fricker', 2.6026896854443837),\n",
       " ('ultimatum', 2.6026896854443837),\n",
       " ('emil', 2.6026896854443837),\n",
       " ('katsu', 2.6026896854443837),\n",
       " ('dev', 2.6026896854443837),\n",
       " ('broadbent', 2.5649493574615367),\n",
       " ('riget', 2.5649493574615367),\n",
       " ('brownstone', 2.5649493574615367),\n",
       " ('holloway', 2.5649493574615367),\n",
       " ('saura', 2.5649493574615367),\n",
       " ('sputnik', 2.5649493574615367),\n",
       " ('girotti', 2.5649493574615367),\n",
       " ('freebird', 2.5649493574615367),\n",
       " ('cynics', 2.5649493574615367),\n",
       " ('boop', 2.5649493574615367),\n",
       " ('dola', 2.5649493574615367),\n",
       " ('tissues', 2.5649493574615367),\n",
       " ('gabriella', 2.5649493574615367),\n",
       " ('mclaglen', 2.5649493574615367),\n",
       " ('mcdoakes', 2.5649493574615367),\n",
       " ('choi', 2.5649493574615367),\n",
       " ('oakie', 2.5649493574615367),\n",
       " ('odysseus', 2.5649493574615367),\n",
       " ('nibelungen', 2.5649493574615367),\n",
       " ('guadalcanal', 2.5649493574615367),\n",
       " ('autograph', 2.5649493574615367),\n",
       " ('yelnats', 2.5649493574615367),\n",
       " ('chavo', 2.5649493574615367),\n",
       " ('bouvier', 2.5649493574615367),\n",
       " ('erendira', 2.5649493574615367),\n",
       " ('unsung', 2.5649493574615367),\n",
       " ('sugiyama', 2.5649493574615367),\n",
       " ('bahrain', 2.5649493574615367),\n",
       " ('luchino', 2.5649493574615367),\n",
       " ('adjani', 2.5649493574615367),\n",
       " ('goring', 2.5649493574615367),\n",
       " ('eglantine', 2.5649493574615367),\n",
       " ('corbett', 2.5494451709255714),\n",
       " ('gandhi', 2.5389738710582761),\n",
       " ('sammi', 2.5389738710582761),\n",
       " ('abu', 2.5389738710582761),\n",
       " ('zu', 2.5389738710582761),\n",
       " ('warhols', 2.5389738710582761),\n",
       " ('delightfully', 2.5257286443082556),\n",
       " ('sirk', 2.5199979695992702),\n",
       " ('rosenstrasse', 2.5123056239761148),\n",
       " ('creasy', 2.5055259369907361),\n",
       " ('braveheart', 2.5014359517392109),\n",
       " ('yuzna', 2.4849066497880004),\n",
       " ('bressart', 2.4849066497880004),\n",
       " ('lassalle', 2.4849066497880004),\n",
       " ('strindberg', 2.4849066497880004),\n",
       " ('cacoyannis', 2.4849066497880004),\n",
       " ('vierde', 2.4849066497880004),\n",
       " ('mcphillip', 2.4849066497880004),\n",
       " ('lupino', 2.4849066497880004),\n",
       " ('santiago', 2.4849066497880004),\n",
       " ('natures', 2.4849066497880004),\n",
       " ('blackadder', 2.4849066497880004),\n",
       " ('barrister', 2.4849066497880004),\n",
       " ('luger', 2.4849066497880004),\n",
       " ('attila', 2.4849066497880004),\n",
       " ('hanlon', 2.4849066497880004),\n",
       " ('gallico', 2.4849066497880004),\n",
       " ('unpretentious', 2.4849066497880004),\n",
       " ('kazan', 2.4849066497880004),\n",
       " ('binder', 2.4849066497880004),\n",
       " ('poonam', 2.4849066497880004),\n",
       " ('trenholm', 2.4849066497880004),\n",
       " ('lafitte', 2.4849066497880004),\n",
       " ('hayao', 2.4849066497880004),\n",
       " ('holodeck', 2.4849066497880004),\n",
       " ('bischoff', 2.4849066497880004),\n",
       " ('parminder', 2.4849066497880004),\n",
       " ('perdition', 2.4849066497880004),\n",
       " ('presque', 2.4849066497880004),\n",
       " ('bouzaglo', 2.4849066497880004),\n",
       " ('talos', 2.4849066497880004),\n",
       " ('euripides', 2.4849066497880004),\n",
       " ('herge', 2.4849066497880004),\n",
       " ('hecht', 2.4849066497880004),\n",
       " ('anselmo', 2.4849066497880004),\n",
       " ('rideau', 2.4849066497880004),\n",
       " ('mcadam', 2.4849066497880004),\n",
       " ('killian', 2.4849066497880004),\n",
       " ('yeon', 2.4849066497880004),\n",
       " ('balduin', 2.4849066497880004),\n",
       " ('clutters', 2.4849066497880004),\n",
       " ('vonnegut', 2.4638532405901681),\n",
       " ('taker', 2.456735772821304),\n",
       " ('mccoy', 2.456735772821304),\n",
       " ('flawless', 2.451005098112319),\n",
       " ('othello', 2.4485390056171252),\n",
       " ('judi', 2.4423470353692043),\n",
       " ('godmother', 2.4423470353692043),\n",
       " ('jonestown', 2.4423470353692043),\n",
       " ('natali', 2.4423470353692043),\n",
       " ('mcnally', 2.4423470353692043),\n",
       " ('mahatma', 2.4423470353692043),\n",
       " ('abbey', 2.4423470353692043),\n",
       " ('novak', 2.4361164856185682),\n",
       " ('durbin', 2.4277482359480516),\n",
       " ('christy', 2.4203681286504293),\n",
       " ('cheadle', 2.4159137783010487),\n",
       " ('faust', 2.3978952727983707),\n",
       " ('hoss', 2.3978952727983707),\n",
       " ('rf', 2.3978952727983707),\n",
       " ('soapdish', 2.3978952727983707),\n",
       " ('bruhl', 2.3978952727983707),\n",
       " ('capano', 2.3978952727983707),\n",
       " ('unsurpassed', 2.3978952727983707),\n",
       " ('vadar', 2.3978952727983707),\n",
       " ('lok', 2.3978952727983707),\n",
       " ('gorris', 2.3978952727983707),\n",
       " ('fuller', 2.3978952727983707),\n",
       " ('viennese', 2.3978952727983707),\n",
       " ('waterman', 2.3978952727983707),\n",
       " ('yasmin', 2.3978952727983707),\n",
       " ('sjoman', 2.3978952727983707),\n",
       " ('antonietta', 2.3978952727983707),\n",
       " ('eisenhower', 2.3978952727983707),\n",
       " ('fellowes', 2.3978952727983707),\n",
       " ('nord', 2.3978952727983707),\n",
       " ('huns', 2.3978952727983707),\n",
       " ('pike', 2.3978952727983707),\n",
       " ('nandini', 2.3978952727983707),\n",
       " ('dorsey', 2.3978952727983707),\n",
       " ('dragoon', 2.3978952727983707),\n",
       " ('starewicz', 2.3978952727983707),\n",
       " ('gundams', 2.3978952727983707),\n",
       " ('noriko', 2.3978952727983707),\n",
       " ('showings', 2.3978952727983707),\n",
       " ('zp', 2.3978952727983707),\n",
       " ('radiant', 2.3978952727983707),\n",
       " ('hallen', 2.3978952727983707),\n",
       " ('taoist', 2.3978952727983707),\n",
       " ('regent', 2.3978952727983707),\n",
       " ('arrondissement', 2.3978952727983707),\n",
       " ('wilhelm', 2.3978952727983707),\n",
       " ('leaud', 2.3978952727983707),\n",
       " ('shep', 2.3978952727983707),\n",
       " ('eminent', 2.3978952727983707),\n",
       " ('stitzer', 2.3978952727983707),\n",
       " ('prologues', 2.3978952727983707),\n",
       " ('cookbook', 2.3978952727983707),\n",
       " ('sweetin', 2.3978952727983707),\n",
       " ('dorfman', 2.3978952727983707),\n",
       " ('fanshawe', 2.3978952727983707),\n",
       " ('hanka', 2.3978952727983707),\n",
       " ('stockwell', 2.3978952727983707),\n",
       " ('bazza', 2.3978952727983707),\n",
       " ('bolan', 2.3978952727983707),\n",
       " ('pei', 2.3978952727983707),\n",
       " ('tykwer', 2.3978952727983707),\n",
       " ('faultless', 2.3978952727983707),\n",
       " ('nyqvist', 2.3978952727983707),\n",
       " ('filone', 2.3978952727983707),\n",
       " ('muska', 2.3978952727983707),\n",
       " ('woronov', 2.3978952727983707),\n",
       " ('spacecamp', 2.3978952727983707),\n",
       " ('tetsur', 2.3978952727983707),\n",
       " ('atul', 2.3978952727983707),\n",
       " ('cannavale', 2.3978952727983707),\n",
       " ('bartel', 2.3978952727983707),\n",
       " ('thursby', 2.3978952727983707),\n",
       " ('rawhide', 2.3978952727983707),\n",
       " ('ramones', 2.3978952727983707),\n",
       " ('seine', 2.3978952727983707),\n",
       " ('gracia', 2.3978952727983707),\n",
       " ('pita', 2.367123614131617),\n",
       " ('cal', 2.3608540011180215),\n",
       " ('schmid', 2.3513752571634776),\n",
       " ('marylee', 2.3513752571634776),\n",
       " ('oberon', 2.3513752571634776),\n",
       " ('johnston', 2.3513752571634776),\n",
       " ('brisson', 2.3513752571634776),\n",
       " ('reservations', 2.3513752571634776),\n",
       " ('elsa', 2.3513752571634776),\n",
       " ('winchester', 2.3353749158170367),\n",
       " ('jabba', 2.3272777055844172),\n",
       " ('chamberlain', 2.3191143949452564),\n",
       " ('zechs', 2.3025850929940459),\n",
       " ('ankush', 2.3025850929940459),\n",
       " ('rollicking', 2.3025850929940459),\n",
       " ('levitt', 2.3025850929940459),\n",
       " ('ori', 2.3025850929940459),\n",
       " ('pasdar', 2.3025850929940459),\n",
       " ('swashbucklers', 2.3025850929940459),\n",
       " ('quebec', 2.3025850929940459),\n",
       " ('manny', 2.3025850929940459),\n",
       " ('katey', 2.3025850929940459),\n",
       " ('bowser', 2.3025850929940459),\n",
       " ('pquerette', 2.3025850929940459),\n",
       " ('gwizdo', 2.3025850929940459),\n",
       " ('hickcock', 2.3025850929940459),\n",
       " ('unparalleled', 2.3025850929940459),\n",
       " ('randell', 2.3025850929940459),\n",
       " ('waterloo', 2.3025850929940459),\n",
       " ('apatow', 2.3025850929940459),\n",
       " ('hardworking', 2.3025850929940459),\n",
       " ('homeward', 2.3025850929940459),\n",
       " ('shetty', 2.3025850929940459),\n",
       " ('hundstage', 2.3025850929940459),\n",
       " ('strombel', 2.3025850929940459),\n",
       " ('rozsa', 2.3025850929940459),\n",
       " ('calvet', 2.3025850929940459),\n",
       " ('milyang', 2.3025850929940459),\n",
       " ('yam', 2.3025850929940459),\n",
       " ('rhapsody', 2.3025850929940459),\n",
       " ('bannister', 2.3025850929940459),\n",
       " ('separating', 2.3025850929940459),\n",
       " ('parkins', 2.3025850929940459),\n",
       " ('ernesto', 2.3025850929940459),\n",
       " ('kidd', 2.3025850929940459),\n",
       " ('beale', 2.3025850929940459),\n",
       " ('lifshitz', 2.3025850929940459),\n",
       " ('poldi', 2.3025850929940459),\n",
       " ('arzenta', 2.3025850929940459),\n",
       " ('manfred', 2.3025850929940459),\n",
       " ('fps', 2.3025850929940459),\n",
       " ('sikes', 2.3025850929940459),\n",
       " ('tollinger', 2.3025850929940459),\n",
       " ('pneumonic', 2.3025850929940459),\n",
       " ('jermaine', 2.3025850929940459),\n",
       " ('ghibli', 2.3025850929940459),\n",
       " ('girlfight', 2.3025850929940459),\n",
       " ('rudyard', 2.3025850929940459),\n",
       " ('spanky', 2.3025850929940459),\n",
       " ('coop', 2.3025850929940459),\n",
       " ('gervais', 2.3025850929940459),\n",
       " ('chamberlains', 2.3025850929940459),\n",
       " ('wuhl', 2.3025850929940459),\n",
       " ('maetel', 2.3025850929940459),\n",
       " ('athens', 2.3025850929940459),\n",
       " ('waxworks', 2.3025850929940459),\n",
       " ('topher', 2.3025850929940459),\n",
       " ('ungar', 2.3025850929940459),\n",
       " ('gilberte', 2.3025850929940459),\n",
       " ('floraine', 2.3025850929940459),\n",
       " ('ginty', 2.3025850929940459),\n",
       " ('hisaishi', 2.3025850929940459),\n",
       " ('ref', 2.3025850929940459),\n",
       " ('miou', 2.3025850929940459),\n",
       " ('jud', 2.3025850929940459),\n",
       " ('montrose', 2.3025850929940459),\n",
       " ('flamenco', 2.3025850929940459),\n",
       " ('colman', 2.3025850929940459),\n",
       " ('overpopulation', 2.3025850929940459),\n",
       " ('virile', 2.3025850929940459),\n",
       " ('rea', 2.3025850929940459),\n",
       " ('broinowski', 2.3025850929940459),\n",
       " ('margo', 2.3025850929940459),\n",
       " ('strickland', 2.3025850929940459),\n",
       " ('haines', 2.3025850929940459),\n",
       " ('mariner', 2.3025850929940459),\n",
       " ('trnka', 2.3025850929940459),\n",
       " ('philipps', 2.3025850929940459),\n",
       " ('thornway', 2.3025850929940459),\n",
       " ('tigerland', 2.3025850929940459),\n",
       " ('tam', 2.3025850929940459),\n",
       " ('prag', 2.3025850929940459),\n",
       " ('hemo', 2.3025850929940459),\n",
       " ('portobello', 2.3025850929940459),\n",
       " ('restoring', 2.3025850929940459),\n",
       " ('carla', 2.2772672850097559),\n",
       " ('fagin', 2.2772672850097559),\n",
       " ('hanzo', 2.2772672850097559),\n",
       " ('aunts', 2.2686835413183641),\n",
       " ('olympia', 2.2686835413183641),\n",
       " ('sullavan', 2.2686835413183641),\n",
       " ('sabrina', 2.2643638801738479),\n",
       " ('superbly', 2.2600254785752498),\n",
       " ('elia', 2.2512917986064953),\n",
       " ('aborigines', 2.2512917986064953),\n",
       " ('hickok', 2.2512917986064953),\n",
       " ('zentropa', 2.2512917986064953),\n",
       " ('peters', 2.2512917986064953),\n",
       " ('linklater', 2.2512917986064953),\n",
       " ('gigi', 2.2512917986064953),\n",
       " ('tully', 2.2512917986064953),\n",
       " ('mccartney', 2.2512917986064953),\n",
       " ('beetle', 2.2512917986064953),\n",
       " ('sweetly', 2.2512917986064953),\n",
       " ('northam', 2.2407096892759584),\n",
       " ('tomlinson', 2.2335922215070942),\n",
       " ('tenant', 2.2284771208403238),\n",
       " ('influential', 2.224623551524334),\n",
       " ('kalifornia', 2.224623551524334),\n",
       " ('stardust', 2.2192034840549946),\n",
       " ('kinnear', 2.2155737160044158),\n",
       " ('madhubala', 2.1972245773362196),\n",
       " ('mes', 2.1972245773362196),\n",
       " ('magnificently', 2.1972245773362196),\n",
       " ('sumpter', 2.1972245773362196),\n",
       " ('tronje', 2.1972245773362196),\n",
       " ('herr', 2.1972245773362196),\n",
       " ('vertigo', 2.1972245773362196),\n",
       " ('szifron', 2.1972245773362196),\n",
       " ('proust', 2.1972245773362196),\n",
       " ('jeffs', 2.1972245773362196),\n",
       " ('xu', 2.1972245773362196),\n",
       " ('hom', 2.1972245773362196),\n",
       " ('doubtlessly', 2.1972245773362196),\n",
       " ('coulardeau', 2.1972245773362196),\n",
       " ('dumbrille', 2.1972245773362196),\n",
       " ('faubourg', 2.1972245773362196),\n",
       " ('vernacular', 2.1972245773362196),\n",
       " ('qualifying', 2.1972245773362196),\n",
       " ('squire', 2.1972245773362196),\n",
       " ('frulein', 2.1972245773362196),\n",
       " ('bizet', 2.1972245773362196),\n",
       " ('feroz', 2.1972245773362196),\n",
       " ('lebrun', 2.1972245773362196),\n",
       " ('timone', 2.1972245773362196),\n",
       " ('cruiserweight', 2.1972245773362196),\n",
       " ('gulpilil', 2.1972245773362196),\n",
       " ('quartier', 2.1972245773362196),\n",
       " ('treasured', 2.1972245773362196),\n",
       " ('futures', 2.1972245773362196),\n",
       " ('machaty', 2.1972245773362196),\n",
       " ('brimmer', 2.1972245773362196),\n",
       " ('lata', 2.1972245773362196),\n",
       " ('smita', 2.1972245773362196),\n",
       " ('forged', 2.1972245773362196),\n",
       " ('faceted', 2.1972245773362196),\n",
       " ('banjos', 2.1972245773362196),\n",
       " ('giuseppe', 2.1972245773362196),\n",
       " ('loners', 2.1972245773362196),\n",
       " ('benet', 2.1972245773362196),\n",
       " ('peepers', 2.1972245773362196),\n",
       " ('ferland', 2.1972245773362196),\n",
       " ('peretti', 2.1972245773362196),\n",
       " ('scarwid', 2.1972245773362196),\n",
       " ('managers', 2.1972245773362196),\n",
       " ('kleinman', 2.1972245773362196),\n",
       " ('pedicab', 2.1972245773362196),\n",
       " ('stig', 2.1972245773362196),\n",
       " ('guignol', 2.1972245773362196),\n",
       " ('hatcher', 2.1972245773362196),\n",
       " ('raines', 2.1972245773362196),\n",
       " ('livesey', 2.1972245773362196),\n",
       " ('sachar', 2.1972245773362196),\n",
       " ('cassio', 2.1972245773362196),\n",
       " ('lawton', 2.1972245773362196),\n",
       " ('dickey', 2.1972245773362196),\n",
       " ('grayce', 2.1972245773362196),\n",
       " ('nwh', 2.1972245773362196),\n",
       " ('montmartre', 2.1972245773362196),\n",
       " ('transports', 2.1972245773362196),\n",
       " ('macchesney', 2.1972245773362196),\n",
       " ('hazlehurst', 2.1972245773362196),\n",
       " ('tuileries', 2.1972245773362196),\n",
       " ('luque', 2.1972245773362196),\n",
       " ('complement', 2.1972245773362196),\n",
       " ('salazar', 2.1972245773362196),\n",
       " ('louella', 2.1972245773362196),\n",
       " ('hermione', 2.1972245773362196),\n",
       " ('romantics', 2.1972245773362196),\n",
       " ('raoul', 2.1972245773362196),\n",
       " ('tatsuhito', 2.1972245773362196),\n",
       " ('imperioli', 2.1972245773362196),\n",
       " ('marchand', 2.1972245773362196),\n",
       " ('maxx', 2.1972245773362196),\n",
       " ('lago', 2.1972245773362196),\n",
       " ('bte', 2.1972245773362196),\n",
       " ('informant', 2.1972245773362196),\n",
       " ('omnibus', 2.1972245773362196),\n",
       " ('vue', 2.1972245773362196),\n",
       " ('torrens', 2.1972245773362196),\n",
       " ('verger', 2.1972245773362196),\n",
       " ('hickory', 2.1972245773362196),\n",
       " ('gespenster', 2.1972245773362196),\n",
       " ('finely', 2.1690537003695232),\n",
       " ('stevenson', 2.1690537003695232),\n",
       " ('cheung', 2.1690537003695232),\n",
       " ('georges', 2.1649637151179979),\n",
       " ('marvelously', 2.1594842493533721),\n",
       " ('rukh', 2.1594842493533721),\n",
       " ('iris', 2.1594842493533721),\n",
       " ('weir', 2.1594842493533721),\n",
       " ('perfection', 2.1594842493533721),\n",
       " ('enchanting', 2.1517622032594619),\n",
       " ('spellbinding', 2.1400661634962708),\n",
       " ('kennel', 2.1400661634962708),\n",
       " ('curr', 2.1400661634962708),\n",
       " ('celine', 2.1400661634962708),\n",
       " ('speakeasy', 2.1400661634962708),\n",
       " ('astaire', 2.1400661634962708),\n",
       " ('pressburger', 2.1400661634962708),\n",
       " ('anand', 2.1400661634962708),\n",
       " ('province', 2.1400661634962708),\n",
       " ('cartwright', 2.1400661634962708),\n",
       " ('leung', 2.1400661634962708),\n",
       " ('dixon', 2.1400661634962708),\n",
       " ('sailing', 2.1400661634962708),\n",
       " ('brodie', 2.1400661634962708),\n",
       " ('mahoney', 2.1400661634962708),\n",
       " ('trier', 2.1316272948504063),\n",
       " ('pecker', 2.1282317058492679),\n",
       " ('wodehouse', 2.120263536200091),\n",
       " ('miyazaki', 2.120263536200091),\n",
       " ('hark', 2.1102132003465894),\n",
       " ('marjorie', 2.1102132003465894),\n",
       " ('darius', 2.1102132003465894),\n",
       " ('vulnerability', 2.1102132003465894),\n",
       " ('devotion', 2.1041341542702074),\n",
       " ('gades', 2.0794415416798357),\n",
       " ('eon', 2.0794415416798357),\n",
       " ('revere', 2.0794415416798357),\n",
       " ('nekron', 2.0794415416798357),\n",
       " ('hennessy', 2.0794415416798357),\n",
       " ('pleasence', 2.0794415416798357),\n",
       " ('zhu', 2.0794415416798357),\n",
       " ('tiempo', 2.0794415416798357),\n",
       " ('malone', 2.0794415416798357),\n",
       " ('sagal', 2.0794415416798357),\n",
       " ('alekos', 2.0794415416798357),\n",
       " ('nihalani', 2.0794415416798357),\n",
       " ('shinjuku', 2.0794415416798357),\n",
       " ('windmill', 2.0794415416798357),\n",
       " ('bragana', 2.0794415416798357),\n",
       " ('kaakha', 2.0794415416798357),\n",
       " ('cristy', 2.0794415416798357),\n",
       " ('touchingly', 2.0794415416798357),\n",
       " ('kangwon', 2.0794415416798357),\n",
       " ('trampa', 2.0794415416798357),\n",
       " ('tieh', 2.0794415416798357),\n",
       " ('ashura', 2.0794415416798357),\n",
       " ('tingler', 2.0794415416798357),\n",
       " ('cinematographers', 2.0794415416798357),\n",
       " ('nyree', 2.0794415416798357),\n",
       " ('patil', 2.0794415416798357),\n",
       " ('saks', 2.0794415416798357),\n",
       " ('intimidated', 2.0794415416798357),\n",
       " ('rostov', 2.0794415416798357),\n",
       " ('loomis', 2.0794415416798357),\n",
       " ('kopins', 2.0794415416798357),\n",
       " ('johanson', 2.0794415416798357),\n",
       " ('mores', 2.0794415416798357),\n",
       " ('jafri', 2.0794415416798357),\n",
       " ('yin', 2.0794415416798357),\n",
       " ('lupin', 2.0794415416798357),\n",
       " ('euthanasia', 2.0794415416798357),\n",
       " ('peaches', 2.0794415416798357),\n",
       " ('herg', 2.0794415416798357),\n",
       " ('genesis', 2.0794415416798357),\n",
       " ('basra', 2.0794415416798357),\n",
       " ('mathis', 2.0794415416798357),\n",
       " ('gainax', 2.0794415416798357),\n",
       " ('boothe', 2.0794415416798357),\n",
       " ('footlight', 2.0794415416798357),\n",
       " ('kemble', 2.0794415416798357),\n",
       " ('malleson', 2.0794415416798357),\n",
       " ('yamamoto', 2.0794415416798357),\n",
       " ('mnm', 2.0794415416798357),\n",
       " ('janeway', 2.0794415416798357),\n",
       " ('versailles', 2.0794415416798357),\n",
       " ('duryea', 2.0794415416798357),\n",
       " ('interweaving', 2.0794415416798357),\n",
       " ('grot', 2.0794415416798357),\n",
       " ('forge', 2.0794415416798357),\n",
       " ('locataire', 2.0794415416798357),\n",
       " ('munshi', 2.0794415416798357),\n",
       " ('cabell', 2.0794415416798357),\n",
       " ('improvising', 2.0794415416798357),\n",
       " ('micheaux', 2.0794415416798357),\n",
       " ('amplified', 2.0794415416798357),\n",
       " ('thuggee', 2.0794415416798357),\n",
       " ('fastway', 2.0794415416798357),\n",
       " ('genma', 2.0794415416798357),\n",
       " ('romy', 2.0794415416798357),\n",
       " ('yang', 2.0794415416798357),\n",
       " ('disarming', 2.0794415416798357),\n",
       " ('salva', 2.0794415416798357),\n",
       " ('meyerling', 2.0794415416798357),\n",
       " ('goodrich', 2.0794415416798357),\n",
       " ('benoit', 2.0794415416798357),\n",
       " ('boundless', 2.0794415416798357),\n",
       " ('abbot', 2.0794415416798357),\n",
       " ('alta', 2.0794415416798357),\n",
       " ('ciannelli', 2.0794415416798357),\n",
       " ('hohl', 2.0794415416798357),\n",
       " ('bedchamber', 2.0794415416798357),\n",
       " ('littlefield', 2.0794415416798357),\n",
       " ('bakersfield', 2.0794415416798357),\n",
       " ('pernell', 2.0794415416798357),\n",
       " ('tackled', 2.0794415416798357),\n",
       " ('unafraid', 2.0794415416798357),\n",
       " ('ryker', 2.0794415416798357),\n",
       " ('brommel', 2.0794415416798357),\n",
       " ('lodoss', 2.0794415416798357),\n",
       " ('apophis', 2.0794415416798357),\n",
       " ('ezra', 2.0794415416798357),\n",
       " ('valientes', 2.0794415416798357),\n",
       " ('sharpshooter', 2.0794415416798357),\n",
       " ('braun', 2.0794415416798357),\n",
       " ('marais', 2.0794415416798357),\n",
       " ('stettner', 2.0794415416798357),\n",
       " ('salles', 2.0794415416798357),\n",
       " ('carax', 2.0794415416798357),\n",
       " ('taj', 2.0794415416798357),\n",
       " ('jodelle', 2.0794415416798357),\n",
       " ('awsome', 2.0794415416798357),\n",
       " ('shinae', 2.0794415416798357),\n",
       " ('thrive', 2.0794415416798357),\n",
       " ('devos', 2.0794415416798357),\n",
       " ('captivates', 2.0794415416798357),\n",
       " ('bromwell', 2.0794415416798357),\n",
       " ('etzel', 2.0794415416798357),\n",
       " ('benito', 2.0794415416798357),\n",
       " ('lachaise', 2.0794415416798357),\n",
       " ('zorak', 2.0794415416798357),\n",
       " ('iago', 2.0794415416798357),\n",
       " ('dimaggio', 2.0794415416798357),\n",
       " ('stphane', 2.0794415416798357),\n",
       " ('technicolour', 2.0794415416798357),\n",
       " ('hiralal', 2.0794415416798357),\n",
       " ('queenie', 2.0794415416798357),\n",
       " ('catiii', 2.0794415416798357),\n",
       " ('valseuses', 2.0794415416798357),\n",
       " ('sandhya', 2.0794415416798357),\n",
       " ('rambeau', 2.0794415416798357),\n",
       " ('fez', 2.0794415416798357),\n",
       " ('inaccessible', 2.0794415416798357),\n",
       " ('boesman', 2.0794415416798357),\n",
       " ('arias', 2.0794415416798357),\n",
       " ('olan', 2.0794415416798357),\n",
       " ('cheech', 2.0541237336955462),\n",
       " ('askey', 2.0476928433652555),\n",
       " ('mol', 2.0476928433652555),\n",
       " ('fairytale', 2.0476928433652555),\n",
       " ('ella', 2.0476928433652555),\n",
       " ('tel', 2.0476928433652555),\n",
       " ('captures', 2.0386195471595809),\n",
       " ('quotable', 2.0368819272610401),\n",
       " ('edith', 2.0368819272610401),\n",
       " ('duchess', 2.0368819272610401),\n",
       " ('curtiz', 2.0368819272610401),\n",
       " ('tintin', 2.0368819272610401),\n",
       " ('scoop', 2.0314323224934752),\n",
       " ('pickford', 2.0314323224934752),\n",
       " ('voight', 2.0301704926730531),\n",
       " ('wonderfully', 2.0218960560332353),\n",
       " ('tetsuo', 2.0149030205422647),\n",
       " ('richness', 2.0149030205422647),\n",
       " ('broomsticks', 2.0149030205422647),\n",
       " ('accomplishes', 2.0149030205422647),\n",
       " ('goa', 2.0149030205422647),\n",
       " ('laird', 2.0149030205422647),\n",
       " ('stalkers', 2.0149030205422647),\n",
       " ('ephemeral', 2.0149030205422647),\n",
       " ('donnell', 2.0149030205422647),\n",
       " ('sasha', 2.0149030205422647),\n",
       " ('blaine', 2.0149030205422647),\n",
       " ('archibald', 2.0149030205422647),\n",
       " ('saffron', 2.0149030205422647),\n",
       " ('patekar', 2.0149030205422647),\n",
       " ('dearly', 2.0149030205422647),\n",
       " ('silvio', 2.0149030205422647),\n",
       " ('milton', 2.0149030205422647),\n",
       " ('brokedown', 2.0149030205422647),\n",
       " ('stubby', 2.0149030205422647),\n",
       " ('kessler', 2.0149030205422647),\n",
       " ('gilley', 2.0149030205422647),\n",
       " ('zorro', 2.0053335695261141),\n",
       " ('clarence', 2.0014800002101243),\n",
       " ('dandy', 1.9980959022258835),\n",
       " ('pawn', 1.9924301646902061),\n",
       " ('egan', 1.9924301646902061),\n",
       " ('bozz', 1.9924301646902061),\n",
       " ('heartbreaking', 1.9924301646902061),\n",
       " ('pickpocket', 1.9924301646902061),\n",
       " ('frailty', 1.9924301646902061),\n",
       " ('venezuela', 1.9810014688665833),\n",
       " ('powell', 1.9783454248084671),\n",
       " ('wang', 1.9740810260220096),\n",
       " ('brosnan', 1.9547990964725592),\n",
       " ('stairway', 1.9459101490553132),\n",
       " ('hightower', 1.9459101490553132),\n",
       " ('vizier', 1.9459101490553132),\n",
       " ('moronie', 1.9459101490553132),\n",
       " ('scaffolding', 1.9459101490553132),\n",
       " ('ligabue', 1.9459101490553132),\n",
       " ('sissi', 1.9459101490553132),\n",
       " ('kikuno', 1.9459101490553132),\n",
       " ('wardh', 1.9459101490553132),\n",
       " ('mellisa', 1.9459101490553132),\n",
       " ('catscratch', 1.9459101490553132),\n",
       " ('conceptions', 1.9459101490553132),\n",
       " ('kurasawa', 1.9459101490553132),\n",
       " ('ethier', 1.9459101490553132),\n",
       " ('lta', 1.9459101490553132),\n",
       " ('meena', 1.9459101490553132),\n",
       " ('margarethe', 1.9459101490553132),\n",
       " ('hersholt', 1.9459101490553132),\n",
       " ('preity', 1.9459101490553132),\n",
       " ('rataud', 1.9459101490553132),\n",
       " ('krimi', 1.9459101490553132),\n",
       " ('melle', 1.9459101490553132),\n",
       " ('symmetry', 1.9459101490553132),\n",
       " ('gilchrist', 1.9459101490553132),\n",
       " ('novela', 1.9459101490553132),\n",
       " ('pyewacket', 1.9459101490553132),\n",
       " ('brak', 1.9459101490553132),\n",
       " ('evacuee', 1.9459101490553132),\n",
       " ('akane', 1.9459101490553132),\n",
       " ('moroder', 1.9459101490553132),\n",
       " ('ida', 1.9459101490553132),\n",
       " ('borough', 1.9459101490553132),\n",
       " ('mavericks', 1.9459101490553132),\n",
       " ('mcclurg', 1.9459101490553132),\n",
       " ('trusts', 1.9459101490553132),\n",
       " ('canto', 1.9459101490553132),\n",
       " ('trebor', 1.9459101490553132),\n",
       " ('cleanliness', 1.9459101490553132),\n",
       " ('dicken', 1.9459101490553132),\n",
       " ('milder', 1.9459101490553132),\n",
       " ('corrine', 1.9459101490553132),\n",
       " ('suo', 1.9459101490553132),\n",
       " ('ros', 1.9459101490553132),\n",
       " ('stunner', 1.9459101490553132),\n",
       " ('rathbone', 1.9459101490553132),\n",
       " ('wtc', 1.9459101490553132),\n",
       " ('unheralded', 1.9459101490553132),\n",
       " ('steppers', 1.9459101490553132),\n",
       " ('enlisting', 1.9459101490553132),\n",
       " ('ripner', 1.9459101490553132),\n",
       " ('tessari', 1.9459101490553132),\n",
       " ('sirico', 1.9459101490553132),\n",
       " ('citizenx', 1.9459101490553132),\n",
       " ('haughty', 1.9459101490553132),\n",
       " ('morand', 1.9459101490553132),\n",
       " ('busfield', 1.9459101490553132),\n",
       " ('rending', 1.9459101490553132),\n",
       " ('pransky', 1.9459101490553132),\n",
       " ('kurtwood', 1.9459101490553132),\n",
       " ('glaser', 1.9459101490553132),\n",
       " ('evocation', 1.9459101490553132),\n",
       " ('warfield', 1.9459101490553132),\n",
       " ('radioactivity', 1.9459101490553132),\n",
       " ('zeon', 1.9459101490553132),\n",
       " ('rawlins', 1.9459101490553132),\n",
       " ('kaley', 1.9459101490553132),\n",
       " ('amrapurkar', 1.9459101490553132),\n",
       " ('marolla', 1.9459101490553132),\n",
       " ('chalo', 1.9459101490553132),\n",
       " ('cosimo', 1.9459101490553132),\n",
       " ('princeton', 1.9459101490553132),\n",
       " ('attaches', 1.9459101490553132),\n",
       " ('gedde', 1.9459101490553132),\n",
       " ('aames', 1.9459101490553132),\n",
       " ('refreshingly', 1.9459101490553132),\n",
       " ('cratchit', 1.9459101490553132),\n",
       " ('dickory', 1.9459101490553132),\n",
       " ('carmela', 1.9459101490553132),\n",
       " ('unwillingly', 1.9459101490553132),\n",
       " ('overjoyed', 1.9459101490553132),\n",
       " ('kagan', 1.9459101490553132),\n",
       " ('trudi', 1.9459101490553132),\n",
       " ('remer', 1.9459101490553132),\n",
       " ('antonia', 1.9459101490553132),\n",
       " ('junkermann', 1.9459101490553132),\n",
       " ('riedelsheimer', 1.9459101490553132),\n",
       " ('scribe', 1.9459101490553132),\n",
       " ('chambara', 1.9459101490553132),\n",
       " ('castorini', 1.9459101490553132),\n",
       " ('bogarde', 1.9459101490553132),\n",
       " ('solvang', 1.9459101490553132),\n",
       " ('steckert', 1.9459101490553132),\n",
       " ('costly', 1.9459101490553132),\n",
       " ('collinwood', 1.9459101490553132),\n",
       " ('wastrel', 1.9459101490553132),\n",
       " ('persistence', 1.9459101490553132),\n",
       " ('burnford', 1.9459101490553132),\n",
       " ('altair', 1.9459101490553132),\n",
       " ('henderson', 1.9459101490553132),\n",
       " ('personifies', 1.9459101490553132),\n",
       " ('laudenbach', 1.9459101490553132),\n",
       " ('differing', 1.9459101490553132),\n",
       " ('francois', 1.9459101490553132),\n",
       " ('momo', 1.9459101490553132),\n",
       " ('bondi', 1.9459101490553132),\n",
       " ('dauphine', 1.9459101490553132),\n",
       " ('kasturba', 1.9459101490553132),\n",
       " ('passworthy', 1.9459101490553132),\n",
       " ('geoff', 1.9459101490553132),\n",
       " ('singin', 1.9459101490553132),\n",
       " ('oshii', 1.9459101490553132),\n",
       " ('bafta', 1.9459101490553132),\n",
       " ('dormael', 1.9459101490553132),\n",
       " ('kazuhiro', 1.9459101490553132),\n",
       " ('concubine', 1.9459101490553132),\n",
       " ('beban', 1.9459101490553132),\n",
       " ('pepi', 1.9459101490553132),\n",
       " ('beulah', 1.9459101490553132),\n",
       " ('izumo', 1.9459101490553132),\n",
       " ('heyerdahl', 1.9459101490553132),\n",
       " ('douglass', 1.9459101490553132),\n",
       " ('korsmo', 1.9459101490553132),\n",
       " ('loulou', 1.9459101490553132),\n",
       " ('transcends', 1.9459101490553132),\n",
       " ('gordone', 1.9459101490553132),\n",
       " ('pom', 1.9459101490553132),\n",
       " ('wymer', 1.9459101490553132),\n",
       " ('tachiguishi', 1.9459101490553132),\n",
       " ('jouvet', 1.9459101490553132),\n",
       " ('slipper', 1.9459101490553132),\n",
       " ('hawke', 1.9459101490553132),\n",
       " ('carre', 1.9459101490553132),\n",
       " ('cantankerous', 1.9459101490553132),\n",
       " ('dirtiest', 1.9459101490553132),\n",
       " ('superstardom', 1.9459101490553132),\n",
       " ('zapatti', 1.9459101490553132),\n",
       " ('yaara', 1.9459101490553132),\n",
       " ('misha', 1.9459101490553132),\n",
       " ('mcdiarmid', 1.9459101490553132),\n",
       " ('aro', 1.9459101490553132),\n",
       " ('vinson', 1.9459101490553132),\n",
       " ('mala', 1.9459101490553132),\n",
       " ('aurora', 1.9459101490553132),\n",
       " ('truman', 1.9459101490553132),\n",
       " ('roo', 1.9459101490553132),\n",
       " ('edelman', 1.9459101490553132),\n",
       " ('heartbreakingly', 1.9459101490553132),\n",
       " ('melfi', 1.9459101490553132),\n",
       " ('tagge', 1.9459101490553132),\n",
       " ('kel', 1.9459101490553132),\n",
       " ('jaco', 1.9459101490553132),\n",
       " ('enthralling', 1.9459101490553132),\n",
       " ('zeenat', 1.9459101490553132),\n",
       " ('reassuring', 1.9459101490553132),\n",
       " ('zarabeth', 1.9459101490553132),\n",
       " ('mostel', 1.9459101490553132),\n",
       " ('strathairn', 1.9459101490553132),\n",
       " ('konvitz', 1.9459101490553132),\n",
       " ('rapids', 1.9459101490553132),\n",
       " ('tur', 1.9459101490553132),\n",
       " ('reopened', 1.9459101490553132),\n",
       " ('donlan', 1.9459101490553132),\n",
       " ('gingold', 1.9459101490553132),\n",
       " ('argentine', 1.9459101490553132),\n",
       " ('sweid', 1.9459101490553132),\n",
       " ('doghi', 1.9459101490553132),\n",
       " ('guarding', 1.9459101490553132),\n",
       " ('thierry', 1.9459101490553132),\n",
       " ('yamada', 1.9459101490553132),\n",
       " ('eikenberry', 1.9459101490553132),\n",
       " ('hlots', 1.9459101490553132),\n",
       " ('cj', 1.9459101490553132),\n",
       " ('doozy', 1.9459101490553132),\n",
       " ('sable', 1.9459101490553132),\n",
       " ('frills', 1.9459101490553132),\n",
       " ('piaf', 1.9459101490553132),\n",
       " ('nanavati', 1.9459101490553132),\n",
       " ('reckoned', 1.9459101490553132),\n",
       " ('mridul', 1.9459101490553132),\n",
       " ...]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# words most frequently seen in a review with a \"POSITIVE\" label\n",
    "pos_neg_ratios.most_common()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('rohtenburg', -4.6051701859880918),\n",
       " ('seriousuly', -4.6051701859880918),\n",
       " ('hoarsely', -4.6051701859880918),\n",
       " ('preexisting', -4.6051701859880918),\n",
       " ('kinlan', -4.6051701859880918),\n",
       " ('jemima', -4.6051701859880918),\n",
       " ('lunchmeat', -4.6051701859880918),\n",
       " ('usaffe', -4.6051701859880918),\n",
       " ('llbean', -4.6051701859880918),\n",
       " ('occassionaly', -4.6051701859880918),\n",
       " ('conclusiondirector', -4.6051701859880918),\n",
       " ('unimposing', -4.6051701859880918),\n",
       " ('contradictors', -4.6051701859880918),\n",
       " ('sondergaard', -4.6051701859880918),\n",
       " ('kapoors', -4.6051701859880918),\n",
       " ('popularism', -4.6051701859880918),\n",
       " ('madoona', -4.6051701859880918),\n",
       " ('diplomas', -4.6051701859880918),\n",
       " ('zenon', -4.6051701859880918),\n",
       " ('petersson', -4.6051701859880918),\n",
       " ('abrasively', -4.6051701859880918),\n",
       " ('vir', -4.6051701859880918),\n",
       " ('phesht', -4.6051701859880918),\n",
       " ('phht', -4.6051701859880918),\n",
       " ('barrages', -4.6051701859880918),\n",
       " ('ist', -4.6051701859880918),\n",
       " ('imperceptibly', -4.6051701859880918),\n",
       " ('rayguns', -4.6051701859880918),\n",
       " ('esra', -4.6051701859880918),\n",
       " ('sprinkler', -4.6051701859880918)]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# words most frequently seen in a review with a \"NEGATIVE\" label\n",
    "list(reversed(pos_neg_ratios.most_common()))[0:30]\n",
    "\n",
    "# Note: Above is the code Andrew uses in his solution video, \n",
    "#       so we've included it here to avoid confusion.\n",
    "#       If you explore the documentation for the Counter class, \n",
    "#       you will see you could also find the 30 least common\n",
    "#       words like this: pos_neg_ratios.most_common()[:-31:-1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# End of Project 1. \n",
    "## Watch the next video to see Andrew's solution, then continue on to the next lesson.\n",
    "\n",
    "# Transforming Text into Numbers<a id='lesson_3'></a>\n",
    "The cells here include code Andrew shows in the next video. We've included it so you can run the code along with the video without having to type in everything."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiIAAAFKCAYAAAAg+zSAAAAABGdBTUEAALGPC/xhBQAAACBjSFJN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\nAH0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import Image\n",
    "\n",
    "review = \"This was a horrible, terrible movie.\"\n",
    "\n",
    "Image(filename='sentiment_network.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAi4AAAECCAYAAADZzFwPAAAABGdBTUEAALGPC/xhBQAAACBjSFJN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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "review = \"The movie was excellent\"\n",
    "\n",
    "Image(filename='sentiment_network_pos.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 2: Creating the Input/Output Data<a id='project_2'></a>\n",
    "\n",
    "**TODO:** Create a [set](https://docs.python.org/3/tutorial/datastructures.html#sets) named `vocab` that contains every word in the vocabulary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# TODO: Create set named \"vocab\" containing all of the words from all of the reviews\n",
    "vocab = set(total_counts.keys())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to check your vocabulary size. If everything worked correctly, it should print **74074**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "74074\n"
     ]
    }
   ],
   "source": [
    "vocab_size = len(vocab)\n",
    "print(vocab_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Take a look at the following image. It represents the layers of the neural network you'll be building throughout this notebook. `layer_0` is the input layer, `layer_1` is a hidden layer, and `layer_2` is the output layer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhYAAAFOCAYAAADaclTUAAAABGdBTUEAALGPC/xhBQAAACBjSFJN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\nZqAIKAKKgCKgCPQPAv8fPYIkU9UI+YoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import Image\n",
    "Image(filename='sentiment_network_2.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**TODO:** Create a numpy array called `layer_0` and initialize it to all zeros. You will find the [zeros](https://docs.scipy.org/doc/numpy/reference/generated/numpy.zeros.html) function particularly helpful here. Be sure you create `layer_0` as a 2-dimensional matrix with 1 row and `vocab_size` columns. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# TODO: Create layer_0 matrix with dimensions 1 by vocab_size, initially filled with zeros\n",
    "layer_0 = np.zeros((1,vocab_size))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell. It should display `(1, 74074)`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 74074)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "layer_0.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiIAAAFKCAYAAAAg+zSAAAAABGdBTUEAALGPC/xhBQAAACBjSFJN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\nAH0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import Image\n",
    "Image(filename='sentiment_network.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`layer_0` contains one entry for every word in the vocabulary, as shown in the above image. We need to make sure we know the index of each word, so run the following cell to create a lookup table that stores the index of every word."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'': 0,\n",
       " 'raspberries': 24610,\n",
       " 'efx': 68262,\n",
       " 'rujal': 1,\n",
       " 'defenitly': 8,\n",
       " 'bro': 37019,\n",
       " 'clearcut': 4,\n",
       " 'asphyxiated': 5,\n",
       " 'gerda': 6,\n",
       " 'ringling': 7,\n",
       " 'videogames': 9,\n",
       " 'samedi': 52300,\n",
       " 'fx': 10,\n",
       " 'sweetened': 11,\n",
       " 'brokeback': 12200,\n",
       " 'endless': 12,\n",
       " 'swigged': 13,\n",
       " 'aquariums': 14,\n",
       " 'erosive': 3,\n",
       " 'mouthful': 20,\n",
       " 'tricking': 16,\n",
       " 'buoyant': 17,\n",
       " 'hours': 19,\n",
       " 'swinging': 2,\n",
       " 'irrepressible': 51845,\n",
       " 'southerners': 21,\n",
       " 'pterodactyl': 22,\n",
       " 'encountering': 24,\n",
       " 'marxist': 25,\n",
       " 'ryunosuke': 27,\n",
       " 'coastal': 25777,\n",
       " 'schwadel': 28,\n",
       " 'bravi': 29,\n",
       " 'jewelery': 32,\n",
       " 'precursor': 65852,\n",
       " 'sexpot': 71297,\n",
       " 'crawford': 33,\n",
       " 'sullying': 34,\n",
       " 'brenden': 36,\n",
       " 'wetters': 44,\n",
       " 'schrott': 37,\n",
       " 'gallinger': 38,\n",
       " 'mcgoldrick': 39,\n",
       " 'nook': 43,\n",
       " 'twined': 41,\n",
       " 'shoddiness': 42,\n",
       " 'register': 61757,\n",
       " 'growing': 45,\n",
       " 'melville': 46,\n",
       " 'reason': 61758,\n",
       " 'tactics': 47,\n",
       " 'recapture': 48,\n",
       " 'wendel': 49,\n",
       " 'cabal': 50,\n",
       " 'montaged': 51,\n",
       " 'lizie': 52,\n",
       " 'upsides': 55,\n",
       " 'azaria': 54,\n",
       " 'archetypes': 18625,\n",
       " 'journalistic': 56,\n",
       " 'bang': 12210,\n",
       " 'gaspar': 59,\n",
       " 'totality': 60,\n",
       " 'afterall': 61,\n",
       " 'curb': 62,\n",
       " 'filling': 63,\n",
       " 'recogniton': 64,\n",
       " 'assailants': 65,\n",
       " 'haysbert': 66,\n",
       " 'ongoingness': 67,\n",
       " 'yeoh': 49393,\n",
       " 'inish': 35848,\n",
       " 'everytime': 23021,\n",
       " 'gildersneeze': 71,\n",
       " 'banners': 72,\n",
       " 'centennial': 76,\n",
       " 'careen': 74,\n",
       " 'obscenities': 75,\n",
       " 'frau': 49395,\n",
       " 'sense': 77,\n",
       " 'blackmore': 78,\n",
       " 'alert': 79,\n",
       " 'landfill': 66604,\n",
       " 'referees': 81,\n",
       " 'pit': 82,\n",
       " 'veering': 83,\n",
       " 'jariwala': 84,\n",
       " 'strived': 85,\n",
       " 'conditioned': 55444,\n",
       " 'unkindness': 61766,\n",
       " 'posters': 86,\n",
       " 'brotherconflict': 12218,\n",
       " 'guddi': 38973,\n",
       " 'winced': 88,\n",
       " 'wakes': 89,\n",
       " 'ancients': 91,\n",
       " 'malplace': 92,\n",
       " 'arcturus': 93,\n",
       " 'cooley': 49398,\n",
       " 'manifestation': 95,\n",
       " 'painter': 96,\n",
       " 'esquire': 97,\n",
       " 'babson': 98,\n",
       " 'slowmo': 28561,\n",
       " 'brummie': 47973,\n",
       " 'arty': 99,\n",
       " 'toothbrush': 100,\n",
       " 'cunningly': 12222,\n",
       " 'aesthetics': 101,\n",
       " 'tease': 102,\n",
       " 'ggooooodd': 103,\n",
       " 'markell': 104,\n",
       " 'jeremie': 106,\n",
       " 'privately': 108,\n",
       " 'liquor': 109,\n",
       " 'donnelly': 110,\n",
       " 'aroused': 111,\n",
       " 'scientistilona': 18,\n",
       " 'cluny': 113,\n",
       " 'copywrite': 114,\n",
       " 'yodeller': 117,\n",
       " 'coghlan': 116,\n",
       " 'crusading': 118,\n",
       " 'moviephysics': 24629,\n",
       " 'continuance': 27756,\n",
       " 'mcgwire': 120,\n",
       " 'boyle': 121,\n",
       " 'characterisations': 66589,\n",
       " 'banter': 123,\n",
       " 'daley': 124,\n",
       " 'terorism': 125,\n",
       " 'panhallagan': 127,\n",
       " 'ronda': 11677,\n",
       " 'weeklies': 128,\n",
       " 'physicists': 130,\n",
       " 'spellbinding': 131,\n",
       " 'oblique': 136,\n",
       " 'easting': 133,\n",
       " 'lambert': 134,\n",
       " 'instructive': 135,\n",
       " 'ama': 137,\n",
       " 'individualistic': 138,\n",
       " 'excessively': 139,\n",
       " 'tora': 23,\n",
       " 'digi': 51602,\n",
       " 'yummm': 142,\n",
       " 'distinctive': 73952,\n",
       " 'aquilae': 143,\n",
       " 'plausible': 144,\n",
       " 'tennapel': 145,\n",
       " 'grady': 147,\n",
       " 'leviticus': 148,\n",
       " 'wining': 149,\n",
       " 'esthetically': 151,\n",
       " 'encyclopidie': 152,\n",
       " 'hurts': 153,\n",
       " 'minette': 154,\n",
       " 'recitals': 49404,\n",
       " 'amagula': 156,\n",
       " 'efficiency': 24635,\n",
       " 'execration': 158,\n",
       " 'uncontaminated': 49403,\n",
       " 'sprocket': 159,\n",
       " 'victimized': 160,\n",
       " 'peephole': 161,\n",
       " 'tomason': 162,\n",
       " 'underlie': 163,\n",
       " 'caterina': 165,\n",
       " 'smirking': 166,\n",
       " 'charlo': 167,\n",
       " 'interlaces': 49408,\n",
       " 'valdez': 169,\n",
       " 'kagemusha': 34521,\n",
       " 'prefabricated': 170,\n",
       " 'fodder': 171,\n",
       " 'coexist': 172,\n",
       " 'inexcusable': 24638,\n",
       " 'sneakpreview': 173,\n",
       " 'equiptment': 174,\n",
       " 'tetsuo': 26,\n",
       " 'ashenden': 176,\n",
       " 'dbbe': 177,\n",
       " 'gianetto': 178,\n",
       " 'makeupon': 49412,\n",
       " 'offs': 179,\n",
       " 'kiesser': 180,\n",
       " 'physique': 5087,\n",
       " 'ambushes': 68654,\n",
       " 'grubbing': 182,\n",
       " 'silsby': 61781,\n",
       " 'jia': 34629,\n",
       " 'tantalising': 183,\n",
       " 'tin': 184,\n",
       " 'untold': 186,\n",
       " 'steward': 187,\n",
       " 'clive': 188,\n",
       " 'haaaaaaaaaaaaaarr': 190,\n",
       " 'tremor': 191,\n",
       " 'jazmine': 192,\n",
       " 'clastrophobic': 195,\n",
       " 'boomslang': 194,\n",
       " 'despicable': 47268,\n",
       " 'succinctly': 196,\n",
       " 'groundbraking': 52482,\n",
       " 'stairway': 197,\n",
       " 'enunciate': 198,\n",
       " 'sndtrk': 199,\n",
       " 'dietrickson': 31,\n",
       " 'flyfisherman': 200,\n",
       " 'unsteady': 202,\n",
       " 'wiper': 204,\n",
       " 'grinds': 205,\n",
       " 'squirted': 206,\n",
       " 'fuing': 207,\n",
       " 'dally': 208,\n",
       " 'preparations': 209,\n",
       " 'counsil': 211,\n",
       " 'moonwalk': 212,\n",
       " 'joel': 213,\n",
       " 'outwits': 214,\n",
       " 'liberating': 216,\n",
       " 'filmography': 217,\n",
       " 'madelene': 220,\n",
       " 'refering': 49417,\n",
       " 'uematsu': 222,\n",
       " 'extreamely': 223,\n",
       " 'glasses': 227,\n",
       " 'honeycomb': 225,\n",
       " 'whackees': 229,\n",
       " 'deltoro': 231,\n",
       " 'strap': 230,\n",
       " 'distaste': 70213,\n",
       " 'copulating': 234,\n",
       " 'newsman': 35,\n",
       " 'torso': 235,\n",
       " 'haack': 239,\n",
       " 'saugages': 240,\n",
       " 'require': 241,\n",
       " 'bargepoles': 242,\n",
       " 'bartender': 244,\n",
       " 'deon': 245,\n",
       " 'unperceptive': 249,\n",
       " 'melenzana': 247,\n",
       " 'boundaries': 250,\n",
       " 'recordist': 251,\n",
       " 'reconstitution': 252,\n",
       " 'bygone': 254,\n",
       " 'kriemshild': 255,\n",
       " 'miiko': 257,\n",
       " 'frakken': 40,\n",
       " 'abernathy': 261,\n",
       " 'hein': 24651,\n",
       " 'chori': 262,\n",
       " 'shogunate': 263,\n",
       " 'ranked': 73149,\n",
       " 'transplants': 264,\n",
       " 'persuasions': 265,\n",
       " 'aura': 266,\n",
       " 'bambou': 267,\n",
       " 'lickin': 268,\n",
       " 'banned': 269,\n",
       " 'lauren': 270,\n",
       " 'americas': 69240,\n",
       " 'shittier': 271,\n",
       " 'crudy': 276,\n",
       " 'pacifical': 275,\n",
       " 'resets': 274,\n",
       " 'workshops': 24653,\n",
       " 'competitive': 277,\n",
       " 'chivalry': 49426,\n",
       " 'take': 278,\n",
       " 'banzai': 72830,\n",
       " 'saw': 279,\n",
       " 'dalek': 283,\n",
       " 'tohs': 281,\n",
       " 'cowers': 284,\n",
       " 'protests': 46522,\n",
       " 'cctv': 286,\n",
       " 'gladiator': 287,\n",
       " 'wichita': 288,\n",
       " 'burgendy': 289,\n",
       " 'moviebeautiful': 290,\n",
       " 'jussi': 291,\n",
       " 'annexing': 292,\n",
       " 'neise': 293,\n",
       " 'lapdog': 294,\n",
       " 'convida': 295,\n",
       " 'preliminaries': 41122,\n",
       " 'euphemizing': 296,\n",
       " 'gariazzo': 297,\n",
       " 'kringle': 298,\n",
       " 'menari': 37055,\n",
       " 'causeway': 46773,\n",
       " 'seep': 299,\n",
       " 'scuttling': 54926,\n",
       " 'rationed': 305,\n",
       " 'schertler': 301,\n",
       " 'refrained': 306,\n",
       " 'transposal': 307,\n",
       " 'kids': 304,\n",
       " 'clutches': 308,\n",
       " 'pistoleers': 309,\n",
       " 'dummies': 3786,\n",
       " 'tripods': 69455,\n",
       " 'relativity': 36022,\n",
       " 'oozed': 310,\n",
       " 'willingness': 311,\n",
       " 'coughthe': 312,\n",
       " 'frederic': 314,\n",
       " 'aimee': 315,\n",
       " 'expose': 316,\n",
       " 'mitevska': 317,\n",
       " 'contention': 318,\n",
       " 'scalisi': 30032,\n",
       " 'watterman': 319,\n",
       " 'chauffeured': 72010,\n",
       " 'turakistan': 53,\n",
       " 'modernize': 321,\n",
       " 'atalante': 12253,\n",
       " 'barriers': 322,\n",
       " 'tetes': 323,\n",
       " 'rim': 24660,\n",
       " 'crushed': 324,\n",
       " 'tickle': 325,\n",
       " 'amphibians': 329,\n",
       " 'ballooned': 328,\n",
       " 'clunes': 330,\n",
       " 'litters': 61807,\n",
       " 'irl': 331,\n",
       " 'andromeda': 57160,\n",
       " 'scrappy': 49432,\n",
       " 'hitchhiking': 332,\n",
       " 'mailings': 333,\n",
       " 'abhishek': 334,\n",
       " 'brownish': 335,\n",
       " 'fishburne': 336,\n",
       " 'spilled': 61809,\n",
       " 'boulange': 337,\n",
       " 'foment': 338,\n",
       " 'appears': 339,\n",
       " 'hana': 340,\n",
       " 'doss': 342,\n",
       " 'bardem': 346,\n",
       " 'tulsa': 345,\n",
       " 'answered': 347,\n",
       " 'schifrin': 348,\n",
       " 'pinfall': 349,\n",
       " 'bayliss': 351,\n",
       " 'advocating': 353,\n",
       " 'aneurysm': 356,\n",
       " 'hoschna': 364,\n",
       " 'potch': 357,\n",
       " 'jackalope': 358,\n",
       " 'andaaz': 363,\n",
       " 'criticising': 361,\n",
       " 'excuse': 362,\n",
       " 'catharine': 55377,\n",
       " 'trudge': 68784,\n",
       " 'robed': 365,\n",
       " 'plodded': 366,\n",
       " 'claymore': 367,\n",
       " 'humanist': 368,\n",
       " 'fnnish': 369,\n",
       " 'marylin': 370,\n",
       " 'spree': 37068,\n",
       " 'reeked': 373,\n",
       " 'applauds': 372,\n",
       " 'rouen': 374,\n",
       " 'metallic': 49441,\n",
       " 'caste': 377,\n",
       " 'mde': 382,\n",
       " 'outsourcing': 379,\n",
       " 'macclane': 69,\n",
       " 'spectrum': 384,\n",
       " 'economics': 61813,\n",
       " 'gusto': 12260,\n",
       " 'belgrad': 386,\n",
       " 'machete': 387,\n",
       " 'quarreling': 388,\n",
       " 'hear': 389,\n",
       " 'considerations': 38308,\n",
       " 'unkill': 390,\n",
       " 'babysitters': 391,\n",
       " 'raged': 392,\n",
       " 'strives': 70354,\n",
       " 'peppers': 68,\n",
       " 'implanting': 394,\n",
       " 'princesses': 395,\n",
       " 'vestment': 396,\n",
       " 'lanka': 397,\n",
       " 'rydell': 398,\n",
       " 'bertie': 399,\n",
       " 'barnabus': 400,\n",
       " 'denzil': 404,\n",
       " 'granter': 405,\n",
       " 'supermarkets': 406,\n",
       " 'castlebeck': 53443,\n",
       " 'innovatory': 408,\n",
       " 'format': 409,\n",
       " 'bustin': 410,\n",
       " 'meandered': 411,\n",
       " 'stout': 48127,\n",
       " 'ebay': 412,\n",
       " 'gopal': 49447,\n",
       " 'yumi': 415,\n",
       " 'chihuahuas': 24675,\n",
       " 'airheads': 417,\n",
       " 'nog': 418,\n",
       " 'gentiles': 419,\n",
       " 'abuser': 24676,\n",
       " 'chooses': 67747,\n",
       " 'adoration': 420,\n",
       " 'mighty': 422,\n",
       " 'boringlane': 424,\n",
       " 'darwinianed': 425,\n",
       " 'interfere': 70,\n",
       " 'ruppert': 429,\n",
       " 'marquand': 428,\n",
       " 'undynamic': 12268,\n",
       " 'wasting': 6830,\n",
       " 'tristain': 37077,\n",
       " 'subsidize': 16436,\n",
       " 'jada': 431,\n",
       " 'winning': 73681,\n",
       " 'trekkin': 433,\n",
       " 'gaspingly': 80,\n",
       " 'airships': 434,\n",
       " 'insecurity': 13433,\n",
       " 'tad': 435,\n",
       " 'algeria': 438,\n",
       " 'misdemeanor': 437,\n",
       " 'susanah': 439,\n",
       " 'dissonance': 38224,\n",
       " 'thinner': 26251,\n",
       " 'skinless': 440,\n",
       " 'lm': 441,\n",
       " 'lures': 442,\n",
       " 'dreamcast': 11772,\n",
       " 'promoters': 443,\n",
       " 'rosenstrae': 444,\n",
       " 'rocaille': 445,\n",
       " 'pets': 446,\n",
       " 'overdoing': 447,\n",
       " 'mooning': 449,\n",
       " 'fkers': 454,\n",
       " 'lockjaw': 451,\n",
       " 'jerrine': 452,\n",
       " 'indigo': 453,\n",
       " 'popculture': 455,\n",
       " 'segway': 51179,\n",
       " 'wrongheaded': 457,\n",
       " 'federline': 459,\n",
       " 'trimming': 460,\n",
       " 'hindered': 463,\n",
       " 'prospering': 462,\n",
       " 'retool': 465,\n",
       " 'quietness': 466,\n",
       " 'parents': 467,\n",
       " 'thinkers': 475,\n",
       " 'carlottai': 469,\n",
       " 'albas': 479,\n",
       " 'gainsborough': 52406,\n",
       " 'chothes': 472,\n",
       " 'gnawing': 476,\n",
       " 'cartoonishness': 474,\n",
       " 'pleasantvillesque': 24685,\n",
       " 'champagne': 477,\n",
       " 'biery': 27959,\n",
       " 'tolkin': 480,\n",
       " 'cillian': 72727,\n",
       " 'ffs': 482,\n",
       " 'diluted': 483,\n",
       " 'masquerading': 49459,\n",
       " 'transgenderism': 486,\n",
       " 'nonsenses': 485,\n",
       " 'papery': 487,\n",
       " 'hardwood': 488,\n",
       " 'teenkill': 18869,\n",
       " 'holliman': 489,\n",
       " 'decommissioned': 24686,\n",
       " 'seek': 12277,\n",
       " 'tooled': 491,\n",
       " 'collinson': 493,\n",
       " 'telekinesis': 494,\n",
       " 'detox': 495,\n",
       " 'hillbilles': 496,\n",
       " 'interminable': 497,\n",
       " 'jarjar': 53734,\n",
       " 'mohamad': 72014,\n",
       " 'postpone': 499,\n",
       " 'paradox': 500,\n",
       " 'hails': 501,\n",
       " 'intestinal': 24689,\n",
       " 'dorkier': 502,\n",
       " 'exploites': 503,\n",
       " 'refunded': 504,\n",
       " 'marketability': 70290,\n",
       " 'mushrooms': 61832,\n",
       " 'plagiarized': 505,\n",
       " 'instructed': 12282,\n",
       " 'weezil': 506,\n",
       " 'batmantas': 507,\n",
       " 'adult': 508,\n",
       " 'grammy': 510,\n",
       " 'makeshift': 61836,\n",
       " 'lyric': 511,\n",
       " 'fidani': 512,\n",
       " 'crunches': 514,\n",
       " 'previewed': 61837,\n",
       " 'venal': 515,\n",
       " 'substantiated': 516,\n",
       " 'sharpton': 517,\n",
       " 'fos': 24693,\n",
       " 'seeping': 25634,\n",
       " 'chucking': 12286,\n",
       " 'socialist': 518,\n",
       " 'unjustly': 25259,\n",
       " 'consacrates': 87,\n",
       " 'polysyllabic': 54274,\n",
       " 'hieroglyphics': 37921,\n",
       " 'foulness': 71604,\n",
       " 'coronets': 520,\n",
       " 'bite': 521,\n",
       " 'cyrus': 522,\n",
       " 'trope': 524,\n",
       " 'aide': 61838,\n",
       " 'gradual': 12291,\n",
       " 'cathy': 525,\n",
       " 'company': 526,\n",
       " 'judels': 527,\n",
       " 'xvid': 528,\n",
       " 'bladerunner': 529,\n",
       " 'advancing': 61840,\n",
       " 'ct': 530,\n",
       " 'graffitiing': 531,\n",
       " 'resuscitate': 532,\n",
       " 'futurise': 535,\n",
       " 'hoyo': 537,\n",
       " 'unsinkable': 536,\n",
       " 'glamorise': 49470,\n",
       " 'miming': 538,\n",
       " 'ahehehe': 540,\n",
       " 'gweneth': 543,\n",
       " 'wexler': 37103,\n",
       " 'hegalhuzen': 61842,\n",
       " 'benjamenta': 544,\n",
       " 'deux': 37104,\n",
       " 'mcbeal': 545,\n",
       " 'thumps': 546,\n",
       " 'barclay': 547,\n",
       " 'treacher': 548,\n",
       " 'crossface': 12295,\n",
       " 'austeniana': 90,\n",
       " 'chemically': 550,\n",
       " 'villechez': 551,\n",
       " 'pathetic': 552,\n",
       " 'ithe': 553,\n",
       " 'acknowledged': 554,\n",
       " 'styled': 556,\n",
       " 'anoying': 557,\n",
       " 'detonator': 69326,\n",
       " 'slack': 558,\n",
       " 'george': 559,\n",
       " 'sailors': 560,\n",
       " 'sakaki': 562,\n",
       " 'bedding': 563,\n",
       " 'inmate': 564,\n",
       " 'alphaville': 49478,\n",
       " 'cowper': 565,\n",
       " 'riker': 566,\n",
       " 'narratives': 567,\n",
       " 'silvermann': 568,\n",
       " 'oosh': 569,\n",
       " 'irk': 570,\n",
       " 'wader': 571,\n",
       " 'lines': 37110,\n",
       " 'upgrade': 572,\n",
       " 'pipers': 573,\n",
       " 'dangling': 574,\n",
       " 'squirter': 94,\n",
       " 'teller': 577,\n",
       " 'nussbaum': 578,\n",
       " 'commence': 579,\n",
       " 'subtitle': 580,\n",
       " 'diffuses': 581,\n",
       " 'bailout': 583,\n",
       " 'smtm': 584,\n",
       " 'norseman': 585,\n",
       " 'unnesicary': 64870,\n",
       " 'dialoques': 586,\n",
       " 'jik': 587,\n",
       " 'arise': 588,\n",
       " 'schtupping': 589,\n",
       " 'michle': 593,\n",
       " 'tomb': 591,\n",
       " 'gills': 594,\n",
       " 'yeaworth': 595,\n",
       " 'memama': 596,\n",
       " 'lavant': 597,\n",
       " 'cute': 598,\n",
       " 'sarducci': 600,\n",
       " 'veidt': 601,\n",
       " 'psmith': 37115,\n",
       " 'malik': 37200,\n",
       " 'fulltime': 603,\n",
       " 'closures': 604,\n",
       " 'diced': 61408,\n",
       " 'friels': 605,\n",
       " 'vamps': 606,\n",
       " 'aap': 608,\n",
       " 'doqui': 617,\n",
       " 'schwarzenberg': 610,\n",
       " 'mindfuk': 611,\n",
       " 'sharkbait': 613,\n",
       " 'herringbone': 614,\n",
       " 'quran': 616,\n",
       " 'screecher': 70893,\n",
       " 'financing': 618,\n",
       " 'exorcist': 619,\n",
       " 'villaronga': 621,\n",
       " 'combination': 61855,\n",
       " 'resumed': 622,\n",
       " 'perversity': 623,\n",
       " 'poofters': 624,\n",
       " 'sexier': 625,\n",
       " 'badel': 626,\n",
       " 'neutral': 69395,\n",
       " 'dominique': 62648,\n",
       " 'resettled': 628,\n",
       " 'stepdaughter': 629,\n",
       " 'teevee': 630,\n",
       " 'clipper': 636,\n",
       " 'thingee': 634,\n",
       " 'mikhalkov': 635,\n",
       " 'largess': 17367,\n",
       " 'projects': 637,\n",
       " 'decapitate': 638,\n",
       " 'succinct': 642,\n",
       " 'selve': 640,\n",
       " 'paedophile': 643,\n",
       " 'deepest': 644,\n",
       " 'conjunction': 67067,\n",
       " 'limaye': 645,\n",
       " 'famille': 24708,\n",
       " 'unconnected': 646,\n",
       " 'freddy': 647,\n",
       " 'greco': 61861,\n",
       " 'moguls': 68347,\n",
       " 'fungicide': 649,\n",
       " 'lamenting': 650,\n",
       " 'isles': 48272,\n",
       " 'hardgore': 652,\n",
       " 'evigan': 12316,\n",
       " 'devoted': 653,\n",
       " 'condemning': 654,\n",
       " 'aisle': 655,\n",
       " 'doritos': 662,\n",
       " 'guffawing': 658,\n",
       " 'inexorable': 659,\n",
       " 'carnosaurs': 664,\n",
       " 'aredavid': 49494,\n",
       " 'authorised': 665,\n",
       " 'durga': 668,\n",
       " 'retitled': 669,\n",
       " 'feeling': 671,\n",
       " 'fooler': 57386,\n",
       " 'puhleeeeze': 672,\n",
       " 'zandt': 65432,\n",
       " 'adverse': 673,\n",
       " 'blacksnake': 675,\n",
       " 'minoan': 69668,\n",
       " 'likeably': 676,\n",
       " 'freshmen': 677,\n",
       " 'garson': 678,\n",
       " 'duologue': 679,\n",
       " 'fluidity': 680,\n",
       " 'camouflaged': 681,\n",
       " 'kasam': 682,\n",
       " 'intoxicants': 684,\n",
       " 'antique': 49291,\n",
       " 'sangster': 685,\n",
       " 'games': 686,\n",
       " 'leprechaun': 43790,\n",
       " 'jealous': 687,\n",
       " 'valediction': 688,\n",
       " 'supporters': 689,\n",
       " 'linz': 690,\n",
       " 'jeannette': 691,\n",
       " 'florakis': 692,\n",
       " 'pointy': 112,\n",
       " 'frocked': 61867,\n",
       " 'aj': 694,\n",
       " 'rainbow': 695,\n",
       " 'eon': 696,\n",
       " 'sn': 697,\n",
       " 'karn': 698,\n",
       " 'isral': 65542,\n",
       " 'silverware': 701,\n",
       " 'aiieeee': 700,\n",
       " 'reportedly': 115,\n",
       " 'tasks': 73986,\n",
       " 'bacons': 702,\n",
       " 'implicate': 706,\n",
       " 'astonished': 704,\n",
       " 'equally': 705,\n",
       " 'pengy': 707,\n",
       " 'womens': 708,\n",
       " 'durham': 119,\n",
       " 'reprehensible': 710,\n",
       " 'jejune': 711,\n",
       " 'taly': 712,\n",
       " 'virgins': 713,\n",
       " 'commodification': 67084,\n",
       " 'screws': 715,\n",
       " 'ungraspable': 716,\n",
       " 'creditsand': 717,\n",
       " 'hoi': 718,\n",
       " 'desperado': 719,\n",
       " 'bikay': 720,\n",
       " 'zeder': 722,\n",
       " 'belligerent': 723,\n",
       " 'haggis': 724,\n",
       " 'anthropomorphics': 12329,\n",
       " 'peaches': 727,\n",
       " 'fastardization': 731,\n",
       " 'eroticism': 728,\n",
       " 'ravensback': 730,\n",
       " 'calgary': 122,\n",
       " 'toxic': 64447,\n",
       " 'yitzhack': 732,\n",
       " 'precode': 72268,\n",
       " 'mcnairy': 733,\n",
       " 'patni': 61876,\n",
       " 'happening': 49508,\n",
       " 'chowder': 734,\n",
       " 'sue': 735,\n",
       " 'peptides': 737,\n",
       " 'overdrive': 61880,\n",
       " 'unattuned': 738,\n",
       " 'shakespear': 739,\n",
       " 'diry': 740,\n",
       " 'hypotheses': 37138,\n",
       " 'petrielli': 741,\n",
       " 'matters': 742,\n",
       " 'huts': 743,\n",
       " 'dcreasy': 744,\n",
       " 'tenth': 745,\n",
       " 'lahr': 746,\n",
       " 'burlinson': 37140,\n",
       " 'dime': 748,\n",
       " 'sentimentalizing': 749,\n",
       " 'ingrate': 750,\n",
       " 'trixie': 751,\n",
       " 'reset': 129,\n",
       " 'langella': 753,\n",
       " 'servitude': 754,\n",
       " 'brooding': 755,\n",
       " 'naushad': 756,\n",
       " 'goldsworthy': 757,\n",
       " 'term': 10507,\n",
       " 'wha': 760,\n",
       " 'aerosol': 761,\n",
       " 'spawning': 763,\n",
       " 'waistband': 767,\n",
       " 'joslyn': 765,\n",
       " 'morland': 65877,\n",
       " 'writter': 768,\n",
       " 'leftover': 769,\n",
       " 'fail': 770,\n",
       " 'huet': 771,\n",
       " 'blalock': 773,\n",
       " 'bumblers': 774,\n",
       " 'role': 132,\n",
       " 'excerpt': 775,\n",
       " 'genitals': 776,\n",
       " 'sikking': 73010,\n",
       " 'dusky': 778,\n",
       " 'inheritance': 779,\n",
       " 'strindberg': 784,\n",
       " 'adolescence': 782,\n",
       " 'shakespearean': 783,\n",
       " 'leninists': 786,\n",
       " 'preachiness': 24730,\n",
       " 'tackiness': 787,\n",
       " 'retard': 788,\n",
       " 'meyerowitz': 789,\n",
       " 'jackies': 793,\n",
       " 'paddling': 792,\n",
       " 'rectified': 794,\n",
       " 'recap': 57115,\n",
       " 'yugoslav': 796,\n",
       " 'preemptively': 797,\n",
       " 'childlish': 17164,\n",
       " 'kind': 798,\n",
       " 'predecessor': 802,\n",
       " 'solanki': 47422,\n",
       " 'deke': 140,\n",
       " 'tolerably': 803,\n",
       " 'sergej': 27551,\n",
       " 'cadavra': 804,\n",
       " 'makepeace': 805,\n",
       " 'resemble': 806,\n",
       " 'buses': 141,\n",
       " 'khanna': 807,\n",
       " 'hypothesized': 53835,\n",
       " 'heaves': 809,\n",
       " 'alexei': 819,\n",
       " 'stacie': 811,\n",
       " 'tinsletown': 812,\n",
       " 'dazzle': 813,\n",
       " 'contemptuously': 814,\n",
       " 'fossilised': 815,\n",
       " 'manhattan': 816,\n",
       " 'quack': 817,\n",
       " 'laurel': 66063,\n",
       " 'vrit': 820,\n",
       " 'lapsing': 821,\n",
       " 'confined': 67384,\n",
       " 'fanny': 49517,\n",
       " 'henrietta': 54398,\n",
       " 'milking': 823,\n",
       " 'reignites': 824,\n",
       " 'quickliy': 826,\n",
       " 'rallying': 827,\n",
       " 'geysers': 829,\n",
       " 'ts': 830,\n",
       " 'knowledgement': 831,\n",
       " 'hitter': 833,\n",
       " 'gatsby': 834,\n",
       " 'hightower': 835,\n",
       " 'artificially': 61899,\n",
       " 'rawer': 838,\n",
       " 'cinematographed': 24738,\n",
       " 'heusen': 839,\n",
       " 'grift': 840,\n",
       " 'corrupt': 841,\n",
       " 'kylie': 842,\n",
       " 'politique': 843,\n",
       " 'karin': 30853,\n",
       " 'veoh': 12345,\n",
       " 'languishes': 845,\n",
       " 'feasts': 64373,\n",
       " 'harrumphing': 846,\n",
       " 'improvements': 847,\n",
       " 'rambos': 12347,\n",
       " 'crepe': 848,\n",
       " 'hindu': 849,\n",
       " 'wean': 850,\n",
       " 'bewilderedly': 851,\n",
       " 'nemec': 35923,\n",
       " 'antic': 853,\n",
       " 'vaitongi': 854,\n",
       " 'ace': 855,\n",
       " 'deadliest': 856,\n",
       " 'foolhardiness': 857,\n",
       " 'cloyed': 61903,\n",
       " 'geometric': 858,\n",
       " 'kennel': 49526,\n",
       " 'stonewashed': 859,\n",
       " 'limburger': 73996,\n",
       " 'guiry': 861,\n",
       " 'militaristic': 71676,\n",
       " 'peeps': 52670,\n",
       " 'perabo': 862,\n",
       " 'chales': 24744,\n",
       " 'rapprochement': 863,\n",
       " 'commissioner': 864,\n",
       " 'squatters': 24745,\n",
       " 'collusion': 67439,\n",
       " 'ailes': 865,\n",
       " 'assurances': 866,\n",
       " 'protestations': 65931,\n",
       " 'melodious': 49527,\n",
       " 'skeet': 868,\n",
       " 'continuity': 155,\n",
       " 'sodomizing': 869,\n",
       " 'humiliations': 870,\n",
       " 'madperson': 871,\n",
       " 'heartpounding': 876,\n",
       " 'dukakis': 873,\n",
       " 'gjon': 874,\n",
       " 'positronic': 12351,\n",
       " 'couturie': 37159,\n",
       " 'sceptical': 61906,\n",
       " 'damiana': 877,\n",
       " 'potentially': 70512,\n",
       " 'vacuousness': 878,\n",
       " 'eventual': 879,\n",
       " 'acquaint': 157,\n",
       " 'blob': 24751,\n",
       " 'handsome': 880,\n",
       " 'angelopoulos': 881,\n",
       " 'trolls': 65948,\n",
       " 'yearned': 12355,\n",
       " 'kaedin': 884,\n",
       " 'bequest': 883,\n",
       " 'burrier': 61911,\n",
       " 'skipper': 61912,\n",
       " 'obscurity': 886,\n",
       " 'demolishing': 887,\n",
       " 'akai': 890,\n",
       " 'resists': 889,\n",
       " 'aged': 893,\n",
       " 'portentously': 892,\n",
       " 'townie': 49533,\n",
       " 'arctic': 894,\n",
       " 'imprezza': 895,\n",
       " 'goldin': 24754,\n",
       " 'dryfus': 898,\n",
       " 'traumas': 897,\n",
       " 'mirroring': 899,\n",
       " 'richness': 900,\n",
       " 'distributing': 901,\n",
       " 'mellowy': 14267,\n",
       " 'adelle': 905,\n",
       " 'hatreds': 906,\n",
       " 'stagecoach': 907,\n",
       " 'hailed': 908,\n",
       " 'taglines': 909,\n",
       " 'finger': 910,\n",
       " 'fanatasy': 67152,\n",
       " 'swilling': 911,\n",
       " 'interviewed': 37164,\n",
       " 'eileen': 912,\n",
       " 'riotous': 915,\n",
       " 'pickin': 916,\n",
       " 'ta': 917,\n",
       " 'reconnecting': 49537,\n",
       " 'degrade': 918,\n",
       " 'leigh': 164,\n",
       " 'toots': 920,\n",
       " 'deadened': 922,\n",
       " 'politic': 12365,\n",
       " 'campground': 924,\n",
       " 'macbeal': 63541,\n",
       " 'iron': 925,\n",
       " 'laz': 926,\n",
       " 'ruler': 3361,\n",
       " 'grufford': 927,\n",
       " 'vickie': 928,\n",
       " 'seifeld': 929,\n",
       " 'indicates': 930,\n",
       " 'bingel': 168,\n",
       " 'kahlua': 933,\n",
       " 'jaco': 935,\n",
       " 'jesse': 49540,\n",
       " 'tanning': 936,\n",
       " 'sisley': 937,\n",
       " 'yeh': 940,\n",
       " 'duper': 939,\n",
       " 'publicized': 942,\n",
       " 'impressible': 52019,\n",
       " 'boulting': 948,\n",
       " 'amoral': 63005,\n",
       " 'conceivable': 949,\n",
       " 'maharashtra': 950,\n",
       " 'venezuelan': 951,\n",
       " 'senator': 71812,\n",
       " 'blabber': 24760,\n",
       " 'tupinambs': 958,\n",
       " 'stop': 955,\n",
       " 'staten': 956,\n",
       " 'slickness': 957,\n",
       " 'dzundza': 959,\n",
       " 'equitable': 965,\n",
       " 'purnell': 961,\n",
       " 'buts': 962,\n",
       " 'cardz': 963,\n",
       " 'scurries': 964,\n",
       " 'precipice': 966,\n",
       " 'housedress': 967,\n",
       " 'verges': 968,\n",
       " 'victim': 969,\n",
       " 'riveted': 61927,\n",
       " 'treacherously': 979,\n",
       " 'attitude': 971,\n",
       " 'increase': 972,\n",
       " 'gaira': 980,\n",
       " 'linens': 976,\n",
       " 'zuckers': 977,\n",
       " 'clitoris': 68599,\n",
       " 'vanilla': 61929,\n",
       " 'wwaste': 984,\n",
       " 'irregularities': 982,\n",
       " 'intermissions': 983,\n",
       " 'harpsichordist': 985,\n",
       " 'rhys': 986,\n",
       " 'privateer': 987,\n",
       " 'carolingian': 38183,\n",
       " 'byambition': 61932,\n",
       " 'federal': 988,\n",
       " 'sold': 989,\n",
       " 'likens': 990,\n",
       " 'rewarded': 991,\n",
       " 'smith': 992,\n",
       " 'rahad': 993,\n",
       " 'alternative': 175,\n",
       " 'madhubala': 995,\n",
       " 'detracted': 996,\n",
       " 'pouchy': 49553,\n",
       " ...}"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a dictionary of words in the vocabulary mapped to index positions\n",
    "# (to be used in layer_0)\n",
    "word2index = {}\n",
    "for i,word in enumerate(vocab):\n",
    "    word2index[word] = i\n",
    "    \n",
    "# display the map of words to indices\n",
    "word2index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**TODO:**  Complete the implementation of `update_input_layer`. It should count \n",
    "          how many times each word is used in the given review, and then store\n",
    "          those counts at the appropriate indices inside `layer_0`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def update_input_layer(review):\n",
    "    \"\"\" Modify the global layer_0 to represent the vector form of review.\n",
    "    The element at a given index of layer_0 should represent\n",
    "    how many times the given word occurs in the review.\n",
    "    Args:\n",
    "        review(string) - the string of the review\n",
    "    Returns:\n",
    "        None\n",
    "    \"\"\"\n",
    "    global layer_0\n",
    "    # clear out previous state by resetting the layer to be all 0s\n",
    "    layer_0 *= 0\n",
    "    \n",
    "    # TODO: count how many times each word is used in the given review and store the results in layer_0 \n",
    "    for wd in review.split(' '):\n",
    "        layer_0[0][word2index[wd]] += 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to test updating the input layer with the first review. The indices assigned may not be the same as in the solution, but hopefully you'll see some non-zero values in `layer_0`.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 18.,   0.,   0., ...,   0.,   0.,   0.]])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "update_input_layer(reviews[0])\n",
    "layer_0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**TODO:** Complete the implementation of `get_target_for_labels`. It should return `0` or `1`, \n",
    "          depending on whether the given label is `NEGATIVE` or `POSITIVE`, respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_target_for_label(label):\n",
    "    \"\"\"Convert a label to `0` or `1`.\n",
    "    Args:\n",
    "        label(string) - Either \"POSITIVE\" or \"NEGATIVE\".\n",
    "    Returns:\n",
    "        `1` or `0`.\n",
    "    \"\"\"\n",
    "    # TODO: Your code here\n",
    "    if label == \"POSITIVE\":\n",
    "        return 1\n",
    "    else:\n",
    "        return 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following two cells. They should print out`'POSITIVE'` and `1`, respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'POSITIVE'"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_target_for_label(labels[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following two cells. They should print out `'NEGATIVE'` and `0`, respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'NEGATIVE'"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_target_for_label(labels[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# End of Project 2. \n",
    "## Watch the next video to see Andrew's solution, then continue on to the next lesson."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 3: Building a Neural Network<a id='project_3'></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**TODO:** We've included the framework of a class called `SentimentNetork`. Implement all of the items marked `TODO` in the code. These include doing the following:\n",
    "- Create a basic neural network much like the networks you've seen in earlier lessons and in Project 1, with an input layer, a hidden layer, and an output layer. \n",
    "- Do **not** add a non-linearity in the hidden layer. That is, do not use an activation function when calculating the hidden layer outputs.\n",
    "- Re-use the code from earlier in this notebook to create the training data (see `TODO`s in the code)\n",
    "- Implement the `pre_process_data` function to create the vocabulary for our training data generating functions\n",
    "- Ensure `train` trains over the entire corpus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import time\n",
    "import sys\n",
    "import numpy as np\n",
    "\n",
    "# Encapsulate our neural network in a class\n",
    "class SentimentNetwork:\n",
    "    def __init__(self, reviews, labels, hidden_nodes = 10, learning_rate = 0.1):\n",
    "        \"\"\"Create a SentimenNetwork with the given settings\n",
    "        Args:\n",
    "            reviews(list) - List of reviews used for training\n",
    "            labels(list) - List of POSITIVE/NEGATIVE labels associated with the given reviews\n",
    "            hidden_nodes(int) - Number of nodes to create in the hidden layer\n",
    "            learning_rate(float) - Learning rate to use while training\n",
    "        \n",
    "        \"\"\"\n",
    "        # Assign a seed to our random number generator to ensure we get\n",
    "        # reproducable results during development \n",
    "        np.random.seed(1)\n",
    "\n",
    "        # process the reviews and their associated labels so that everything\n",
    "        # is ready for training\n",
    "        self.pre_process_data(reviews, labels)\n",
    "        \n",
    "        # Build the network to have the number of hidden nodes and the learning rate that\n",
    "        # were passed into this initializer. Make the same number of input nodes as\n",
    "        # there are vocabulary words and create a single output node.\n",
    "        self.init_network(len(self.review_vocab),hidden_nodes, 1, learning_rate)\n",
    "\n",
    "    def pre_process_data(self, reviews, labels):\n",
    "        \n",
    "        review_vocab = set()\n",
    "        # By definition, the sets module provides classes for constructing and manipulating unordered collections of unique elements.\n",
    "        \n",
    "        # TODO: populate review_vocab with all of the words in the given reviews\n",
    "        #       Remember to split reviews into individual words \n",
    "        #       using \"split(' ')\" instead of \"split()\".\n",
    "        for thisReview in reviews:\n",
    "            for word in thisReview.split(' '):\n",
    "                review_vocab.add(word)\n",
    "        # Convert the vocabulary set to a list so we can access words via indices\n",
    "        self.review_vocab = list(review_vocab)\n",
    "        \n",
    "        \n",
    "        label_vocab = set()\n",
    "        # TODO: populate label_vocab with all of the words in the given labels.\n",
    "        #       There is no need to split the labels because each one is a single word.\n",
    "        for lbl in labels:\n",
    "            label_vocab.add(lbl)\n",
    "        # Convert the label vocabulary set to a list so we can access labels via indices\n",
    "        self.label_vocab = list(label_vocab)\n",
    "        \n",
    "        \n",
    "        # Store the sizes of the review and label vocabularies.\n",
    "        self.review_vocab_size = len(self.review_vocab)\n",
    "        self.label_vocab_size = len(self.label_vocab)\n",
    "        \n",
    "        \n",
    "        # Create a dictionary of words in the vocabulary mapped to index positions\n",
    "        self.word2index = {}\n",
    "        # TODO: populate self.word2index with indices for all the words in self.review_vocab\n",
    "        #       like you saw earlier in the notebook\n",
    "        for i,word in enumerate(self.review_vocab):\n",
    "            self.word2index[word] = i\n",
    "        \n",
    "        # Create a dictionary of labels mapped to index positions\n",
    "        self.label2index = {}\n",
    "        # TODO: do the same thing you did for self.word2index and self.review_vocab, \n",
    "        #       but for self.label2index and self.label_vocab instead\n",
    "        for i,word in enumerate(self.label_vocab):\n",
    "            self.label2index[word] = i\n",
    "         \n",
    "        \n",
    "    def init_network(self, input_nodes, hidden_nodes, output_nodes, learning_rate):\n",
    "        # Store the number of nodes in input, hidden, and output layers.\n",
    "        self.input_nodes = input_nodes\n",
    "        self.hidden_nodes = hidden_nodes\n",
    "        self.output_nodes = output_nodes\n",
    "\n",
    "        # Store the learning rate\n",
    "        self.learning_rate = learning_rate\n",
    "\n",
    "        # Initialize weights\n",
    "        \n",
    "        # TODO: initialize self.weights_0_1 as a matrix of zeros. These are the weights between\n",
    "        #       the input layer and the hidden layer.\n",
    "        self.weights_0_1 = np.zeros((self.input_nodes,self.hidden_nodes))\n",
    "        # TODO: initialize self.weights_1_2 as a matrix of random values. \n",
    "        #       These are the weights between the hidden layer and the output layer.\n",
    "        #self.weights_1_2 = np.random.rand(self.hidden_nodes, self.output_nodes)\n",
    "        \n",
    "        self.weights_1_2 = np.random.normal(0.0, self.output_nodes**-0.5, \n",
    "                                                (self.hidden_nodes, self.output_nodes))\n",
    "        \n",
    "        # TODO: Create the input layer, a two-dimensional matrix with shape \n",
    "        #       1 x input_nodes, with all values initialized to zero\n",
    "        self.layer_0 = np.zeros((1,input_nodes))\n",
    "    \n",
    "    #---------------------------------------------------------------------------  \n",
    "    def update_input_layer(self,review):\n",
    "        # TODO: You can copy most of the code you wrote for update_input_layer \n",
    "        #       earlier in this notebook. \n",
    "        # clear out previous state by resetting the layer to be all 0s\n",
    "        self.layer_0 *= 0\n",
    "    \n",
    "        # TODO: count how many times each word is used in the given review and store the results in layer_0 \n",
    "        for wd in review.split(' '):\n",
    "            if(wd in self.word2index.keys()):\n",
    "                self.layer_0[0][self.word2index[wd]] += 1\n",
    "            \n",
    "    #---------------------------------------------------------------------------\n",
    "    def get_target_for_label(self,label):\n",
    "        # TODO: Copy the code you wrote for get_target_for_label \n",
    "        #       earlier in this notebook. \n",
    "        if label == \"POSITIVE\":\n",
    "            return 1\n",
    "        else:\n",
    "            return 0\n",
    "        \n",
    "    #---------------------------------------------------------------------------\n",
    "    def sigmoid(self,x):\n",
    "        # TODO: Return the result of calculating the sigmoid activation function\n",
    "        #       shown in the lectures\n",
    "        return (1/(1+np.exp(-x)))\n",
    "    \n",
    "    def sigmoid_output_2_derivative(self,output):\n",
    "        # TODO: Return the derivative of the sigmoid activation function, \n",
    "        #       where \"output\" is the original output from the sigmoid function \n",
    "        return (output * (1-output) )\n",
    "\n",
    "    def train(self, training_reviews, training_labels):\n",
    "        \n",
    "        # make sure out we have a matching number of reviews and labels\n",
    "        assert(len(training_reviews) == len(training_labels))\n",
    "        \n",
    "        # Keep track of correct predictions to display accuracy during training \n",
    "        correct_so_far = 0\n",
    "        \n",
    "        # Remember when we started for printing time statistics\n",
    "        start = time.time()\n",
    "\n",
    "        # loop through all the given reviews and run a forward and backward pass,\n",
    "        # updating weights for every item\n",
    "        for i in range(len(training_reviews)):\n",
    "            \n",
    "            # TODO: Get the next review and its correct label\n",
    "            review = training_reviews[i]\n",
    "            label = training_labels[i]\n",
    "            \n",
    "            # TODO: Implement the forward pass through the network. \n",
    "            #       That means use the given review to update the input layer, \n",
    "            #       then calculate values for the hidden layer,\n",
    "            #       and finally calculate the output layer.\n",
    "            # \n",
    "            #       Do not use an activation function for the hidden layer,\n",
    "            #       but use the sigmoid activation function for the output layer.\n",
    "            self.update_input_layer(review)            \n",
    "            layer_1 = self.layer_0.dot(self.weights_0_1)\n",
    "            layer_2 = self.sigmoid(layer_1.dot(self.weights_1_2))\n",
    "                                    \n",
    "                                    \n",
    "            # TODO: Implement the back propagation pass here. \n",
    "            #       That means calculate the error for the forward pass's prediction\n",
    "            #       and update the weights in the network according to their\n",
    "            #       contributions toward the error, as calculated via the\n",
    "            #       gradient descent and back propagation algorithms you \n",
    "            #       learned in class.\n",
    "\n",
    "            layer_2_error = layer_2 - self.get_target_for_label(label)       # Output layer error == (True label - output)\n",
    "            layer_2_delta = layer_2_error * self.sigmoid_output_2_derivative(layer_2)\n",
    "\n",
    "            layer_1_error = layer_2_delta.dot(self.weights_1_2.T)\n",
    "            layer_1_delta = layer_1_error\n",
    "\n",
    "            self.weights_1_2 -= layer_1.T.dot(layer_2_delta) * self.learning_rate\n",
    "            self.weights_0_1 -= self.layer_0.T.dot(layer_1_delta) * self.learning_rate\n",
    "\n",
    "\n",
    "            \n",
    "            # TODO: Keep track of correct predictions. To determine if the prediction was\n",
    "            #       correct, check that the absolute value of the output error \n",
    "            #       is less than 0.5. If so, add one to the correct_so_far count.\n",
    "            if(layer_2 >= 0.5 and label == 'POSITIVE'):\n",
    "                correct_so_far += 1\n",
    "            elif(layer_2 < 0.5 and label == 'NEGATIVE'):\n",
    "                correct_so_far += 1            \n",
    "            \n",
    "            \n",
    "            # For debug purposes, print out our prediction accuracy and speed \n",
    "            # throughout the training process. \n",
    "\n",
    "            elapsed_time = float(time.time() - start)\n",
    "            reviews_per_second = i / elapsed_time if elapsed_time > 0 else 0\n",
    "            \n",
    "            sys.stdout.write(\"\\rProgress:\" + str(100 * i/float(len(training_reviews)))[:4] \\\n",
    "                             + \"% Speed(reviews/sec):\" + str(reviews_per_second)[0:5] \\\n",
    "                             + \" #Correct:\" + str(correct_so_far) + \" #Trained:\" + str(i+1) \\\n",
    "                             + \" Training Accuracy:\" + str(correct_so_far * 100 / float(i+1))[:4] + \"%\")\n",
    "            if(i % 2500 == 0):\n",
    "                print(\"\")\n",
    "    \n",
    "    def test(self, testing_reviews, testing_labels):\n",
    "        \"\"\"\n",
    "        Attempts to predict the labels for the given testing_reviews,\n",
    "        and uses the test_labels to calculate the accuracy of those predictions.\n",
    "        \"\"\"\n",
    "        \n",
    "        # keep track of how many correct predictions we make\n",
    "        correct = 0\n",
    "\n",
    "        # we'll time how many predictions per second we make\n",
    "        start = time.time()\n",
    "\n",
    "        # Loop through each of the given reviews and call run to predict\n",
    "        # its label. \n",
    "        for i in range(len(testing_reviews)):\n",
    "            pred = self.run(testing_reviews[i])\n",
    "            if(pred == testing_labels[i]):\n",
    "                correct += 1\n",
    "            \n",
    "            # For debug purposes, print out our prediction accuracy and speed \n",
    "            # throughout the prediction process. \n",
    "\n",
    "            elapsed_time = float(time.time() - start)\n",
    "            reviews_per_second = i / elapsed_time if elapsed_time > 0 else 0\n",
    "            \n",
    "            sys.stdout.write(\"\\rProgress:\" + str(100 * i/float(len(testing_reviews)))[:4] \\\n",
    "                             + \"% Speed(reviews/sec):\" + str(reviews_per_second)[0:5] \\\n",
    "                             + \" #Correct:\" + str(correct) + \" #Tested:\" + str(i+1) \\\n",
    "                             + \" Testing Accuracy:\" + str(correct * 100 / float(i+1))[:4] + \"%\")\n",
    "    \n",
    "    def run(self, review):\n",
    "        \"\"\"\n",
    "        Returns a POSITIVE or NEGATIVE prediction for the given review.\n",
    "        \"\"\"\n",
    "        # TODO: Run a forward pass through the network, like you did in the\n",
    "        #       \"train\" function. That means use the given review to \n",
    "        #       update the input layer, then calculate values for the hidden layer,\n",
    "        #       and finally calculate the output layer.\n",
    "        #\n",
    "        #       Note: The review passed into this function for prediction \n",
    "        #             might come from anywhere, so you should convert it \n",
    "        #             to lower case prior to using it.\n",
    "        \n",
    "        # TODO: The output layer should now contain a prediction. \n",
    "        #       Return `POSITIVE` for predictions greater-than-or-equal-to `0.5`, \n",
    "        #       and `NEGATIVE` otherwise.\n",
    "        \n",
    "        self.update_input_layer(review.lower())\n",
    "        layer_1 = self.layer_0.dot(self.weights_0_1)\n",
    "        layer_2 = self.sigmoid(layer_1.dot(self.weights_1_2))\n",
    "        \n",
    "        if(layer_2[0] >= 0.5):\n",
    "            return \"POSITIVE\"\n",
    "        else:\n",
    "            return \"NEGATIVE\"\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to create a `SentimentNetwork` that will train on all but the last 1000 reviews (we're saving those for testing). Here we use a learning rate of `0.1`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mlp = SentimentNetwork(reviews[:-1000],labels[:-1000], learning_rate=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to test the network's performance against the last 1000 reviews (the ones we held out from our training set). \n",
    "\n",
    "**We have not trained the model yet, so the results should be about 50% as it will just be guessing and there are only two possible values to choose from.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:99.9% Speed(reviews/sec):514.2 #Correct:500 #Tested:1000 Testing Accuracy:50.0%"
     ]
    }
   ],
   "source": [
    "mlp.test(reviews[-1000:],labels[-1000:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to actually train the network. During training, it will display the model's accuracy repeatedly as it trains so you can see how well it's doing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:0.0% Speed(reviews/sec):0.0 #Correct:1 #Trained:1 Training Accuracy:100.%\n",
      "Progress:10.4% Speed(reviews/sec):85.47 #Correct:1251 #Trained:2501 Training Accuracy:50.0%\n",
      "Progress:20.8% Speed(reviews/sec):88.53 #Correct:2501 #Trained:5001 Training Accuracy:50.0%\n",
      "Progress:31.2% Speed(reviews/sec):90.14 #Correct:3751 #Trained:7501 Training Accuracy:50.0%\n",
      "Progress:41.6% Speed(reviews/sec):91.06 #Correct:5001 #Trained:10001 Training Accuracy:50.0%\n",
      "Progress:52.0% Speed(reviews/sec):91.50 #Correct:6251 #Trained:12501 Training Accuracy:50.0%\n",
      "Progress:62.5% Speed(reviews/sec):91.76 #Correct:7501 #Trained:15001 Training Accuracy:50.0%\n",
      "Progress:72.9% Speed(reviews/sec):92.02 #Correct:8751 #Trained:17501 Training Accuracy:50.0%\n",
      "Progress:83.3% Speed(reviews/sec):92.01 #Correct:10001 #Trained:20001 Training Accuracy:50.0%\n",
      "Progress:93.7% Speed(reviews/sec):89.37 #Correct:11251 #Trained:22501 Training Accuracy:50.0%\n",
      "Progress:99.9% Speed(reviews/sec):88.23 #Correct:12000 #Trained:24000 Training Accuracy:50.0%"
     ]
    }
   ],
   "source": [
    "mlp.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That most likely didn't train very well. Part of the reason may be because the learning rate is too high. Run the following cell to recreate the network with a smaller learning rate, `0.01`, and then train the new network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:0.0% Speed(reviews/sec):0.0 #Correct:1 #Trained:1 Training Accuracy:100.%\n",
      "Progress:10.4% Speed(reviews/sec):75.40 #Correct:1248 #Trained:2501 Training Accuracy:49.9%\n",
      "Progress:20.8% Speed(reviews/sec):76.39 #Correct:2498 #Trained:5001 Training Accuracy:49.9%\n",
      "Progress:31.2% Speed(reviews/sec):71.88 #Correct:3748 #Trained:7501 Training Accuracy:49.9%\n",
      "Progress:41.6% Speed(reviews/sec):73.32 #Correct:4998 #Trained:10001 Training Accuracy:49.9%\n",
      "Progress:52.0% Speed(reviews/sec):74.13 #Correct:6248 #Trained:12501 Training Accuracy:49.9%\n",
      "Progress:62.5% Speed(reviews/sec):73.62 #Correct:7490 #Trained:15001 Training Accuracy:49.9%\n",
      "Progress:72.9% Speed(reviews/sec):74.09 #Correct:8740 #Trained:17501 Training Accuracy:49.9%\n",
      "Progress:83.3% Speed(reviews/sec):74.86 #Correct:9990 #Trained:20001 Training Accuracy:49.9%\n",
      "Progress:93.7% Speed(reviews/sec):75.60 #Correct:11240 #Trained:22501 Training Accuracy:49.9%\n",
      "Progress:99.9% Speed(reviews/sec):75.96 #Correct:11989 #Trained:24000 Training Accuracy:49.9%"
     ]
    }
   ],
   "source": [
    "mlp = SentimentNetwork(reviews[:-1000],labels[:-1000], learning_rate=0.01)\n",
    "mlp.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That probably wasn't much different. Run the following cell to recreate the network one more time with an even smaller learning rate, `0.001`, and then train the new network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:0.0% Speed(reviews/sec):0.0 #Correct:1 #Trained:1 Training Accuracy:100.%\n",
      "Progress:10.4% Speed(reviews/sec):72.38 #Correct:1255 #Trained:2501 Training Accuracy:50.1%\n",
      "Progress:20.8% Speed(reviews/sec):75.42 #Correct:2611 #Trained:5001 Training Accuracy:52.2%\n",
      "Progress:31.2% Speed(reviews/sec):77.80 #Correct:4071 #Trained:7501 Training Accuracy:54.2%\n",
      "Progress:41.6% Speed(reviews/sec):79.05 #Correct:5626 #Trained:10001 Training Accuracy:56.2%\n",
      "Progress:52.0% Speed(reviews/sec):79.01 #Correct:7101 #Trained:12501 Training Accuracy:56.8%\n",
      "Progress:62.5% Speed(reviews/sec):79.13 #Correct:8671 #Trained:15001 Training Accuracy:57.8%\n",
      "Progress:72.9% Speed(reviews/sec):79.58 #Correct:10319 #Trained:17501 Training Accuracy:58.9%\n",
      "Progress:83.3% Speed(reviews/sec):79.34 #Correct:12002 #Trained:20001 Training Accuracy:60.0%\n",
      "Progress:93.7% Speed(reviews/sec):79.25 #Correct:13684 #Trained:22501 Training Accuracy:60.8%\n",
      "Progress:99.9% Speed(reviews/sec):79.35 #Correct:14713 #Trained:24000 Training Accuracy:61.3%"
     ]
    }
   ],
   "source": [
    "mlp = SentimentNetwork(reviews[:-1000],labels[:-1000], learning_rate=0.001)\n",
    "mlp.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With a learning rate of `0.001`, the network should finall have started to improve during training. It's still not very good, but it shows that this solution has potential. We will improve it in the next lesson."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# End of Project 3. \n",
    "## Watch the next video to see Andrew's solution, then continue on to the next lesson."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Understanding Neural Noise<a id='lesson_4'></a>\n",
    "\n",
    "The following cells include includes the code Andrew shows in the next video. We've included it here so you can run the cells along with the video without having to type in everything."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiIAAAFKCAYAAAAg+zSAAAAABGdBTUEAALGPC/xhBQAAACBjSFJN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\nAH0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import Image\n",
    "Image(filename='sentiment_network.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def update_input_layer(review):\n",
    "    \n",
    "    global layer_0\n",
    "    \n",
    "    # clear out previous state, reset the layer to be all 0s\n",
    "    layer_0 *= 0\n",
    "    for word in review.split(\" \"):\n",
    "        layer_0[0][word2index[word]] += 1\n",
    "\n",
    "update_input_layer(reviews[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 18.,   0.,   0., ...,   0.,   0.,   0.]])"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "layer_0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "review_counter = Counter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for word in reviews[0].split(\" \"):\n",
    "    review_counter[word] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('.', 27),\n",
       " ('', 18),\n",
       " ('the', 9),\n",
       " ('to', 6),\n",
       " ('high', 5),\n",
       " ('i', 5),\n",
       " ('bromwell', 4),\n",
       " ('is', 4),\n",
       " ('of', 4),\n",
       " ('that', 4),\n",
       " ('a', 4),\n",
       " ('teachers', 4),\n",
       " ('school', 2),\n",
       " ('as', 2),\n",
       " ('it', 2),\n",
       " ('me', 2),\n",
       " ('in', 2),\n",
       " ('their', 2),\n",
       " ('students', 2),\n",
       " ('student', 2),\n",
       " ('at', 2),\n",
       " ('my', 2),\n",
       " ('age', 1),\n",
       " ('satire', 1),\n",
       " ('cartoon', 1),\n",
       " ('many', 1),\n",
       " ('pomp', 1),\n",
       " ('all', 1),\n",
       " ('much', 1),\n",
       " ('line', 1),\n",
       " ('reality', 1),\n",
       " ('believe', 1),\n",
       " ('t', 1),\n",
       " ('adults', 1),\n",
       " ('whole', 1),\n",
       " ('which', 1),\n",
       " ('when', 1),\n",
       " ('isn', 1),\n",
       " ('other', 1),\n",
       " ('lead', 1),\n",
       " ('classic', 1),\n",
       " ('burn', 1),\n",
       " ('about', 1),\n",
       " ('time', 1),\n",
       " ('programs', 1),\n",
       " ('sack', 1),\n",
       " ('here', 1),\n",
       " ('recalled', 1),\n",
       " ('think', 1),\n",
       " ('through', 1),\n",
       " ('expect', 1),\n",
       " ('pity', 1),\n",
       " ('schools', 1),\n",
       " ('tried', 1),\n",
       " ('down', 1),\n",
       " ('situation', 1),\n",
       " ('one', 1),\n",
       " ('ran', 1),\n",
       " ('teaching', 1),\n",
       " ('far', 1),\n",
       " ('such', 1),\n",
       " ('survive', 1),\n",
       " ('right', 1),\n",
       " ('profession', 1),\n",
       " ('what', 1),\n",
       " ('episode', 1),\n",
       " ('welcome', 1),\n",
       " ('and', 1),\n",
       " ('see', 1),\n",
       " ('some', 1),\n",
       " ('your', 1),\n",
       " ('repeatedly', 1),\n",
       " ('knew', 1),\n",
       " ('insightful', 1),\n",
       " ('pathetic', 1),\n",
       " ('years', 1),\n",
       " ('immediately', 1),\n",
       " ('life', 1),\n",
       " ('same', 1),\n",
       " ('saw', 1),\n",
       " ('remind', 1),\n",
       " ('scramble', 1),\n",
       " ('can', 1),\n",
       " ('m', 1),\n",
       " ('s', 1),\n",
       " ('who', 1),\n",
       " ('comedy', 1),\n",
       " ('fetched', 1),\n",
       " ('than', 1),\n",
       " ('closer', 1),\n",
       " ('inspector', 1),\n",
       " ('pettiness', 1),\n",
       " ('financially', 1)]"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "review_counter.most_common()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 4: Reducing Noise in Our Input Data<a id='project_4'></a>\n",
    "\n",
    "**TODO:** Attempt to reduce the noise in the input data like Andrew did in the previous video. Specifically, do the following:\n",
    "* Copy the `SentimentNetwork` class you created earlier into the following cell.\n",
    "* Modify `update_input_layer` so it does not count how many times each word is used, but rather just stores whether or not a word was used. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO: -Copy the SentimentNetwork class from Projet 3 lesson\n",
    "#       -Modify it to reduce noise, like in the video \n",
    "# Encapsulate our neural network in a class\n",
    "class SentimentNetwork:\n",
    "    def __init__(self, reviews, labels, hidden_nodes = 10, learning_rate = 0.1):\n",
    "        \"\"\"Create a SentimenNetwork with the given settings\n",
    "        Args:\n",
    "            reviews(list) - List of reviews used for training\n",
    "            labels(list) - List of POSITIVE/NEGATIVE labels associated with the given reviews\n",
    "            hidden_nodes(int) - Number of nodes to create in the hidden layer\n",
    "            learning_rate(float) - Learning rate to use while training\n",
    "        \n",
    "        \"\"\"\n",
    "        # Assign a seed to our random number generator to ensure we get\n",
    "        # reproducable results during development \n",
    "        np.random.seed(1)\n",
    "\n",
    "        # process the reviews and their associated labels so that everything\n",
    "        # is ready for training\n",
    "        self.pre_process_data(reviews, labels)\n",
    "        \n",
    "        # Build the network to have the number of hidden nodes and the learning rate that\n",
    "        # were passed into this initializer. Make the same number of input nodes as\n",
    "        # there are vocabulary words and create a single output node.\n",
    "        self.init_network(len(self.review_vocab),hidden_nodes, 1, learning_rate)\n",
    "\n",
    "    def pre_process_data(self, reviews, labels):\n",
    "        \n",
    "        review_vocab = set()\n",
    "        # By definition, the sets module provides classes for constructing and manipulating unordered collections of unique elements.\n",
    "        \n",
    "        # TODO: populate review_vocab with all of the words in the given reviews\n",
    "        #       Remember to split reviews into individual words \n",
    "        #       using \"split(' ')\" instead of \"split()\".\n",
    "        for thisReview in reviews:\n",
    "            for word in thisReview.split(' '):\n",
    "                review_vocab.add(word)\n",
    "        # Convert the vocabulary set to a list so we can access words via indices\n",
    "        self.review_vocab = list(review_vocab)\n",
    "        \n",
    "        \n",
    "        label_vocab = set()\n",
    "        # TODO: populate label_vocab with all of the words in the given labels.\n",
    "        #       There is no need to split the labels because each one is a single word.\n",
    "        for lbl in labels:\n",
    "            label_vocab.add(lbl)\n",
    "        # Convert the label vocabulary set to a list so we can access labels via indices\n",
    "        self.label_vocab = list(label_vocab)\n",
    "        \n",
    "        \n",
    "        # Store the sizes of the review and label vocabularies.\n",
    "        self.review_vocab_size = len(self.review_vocab)\n",
    "        self.label_vocab_size = len(self.label_vocab)\n",
    "        \n",
    "        \n",
    "        # Create a dictionary of words in the vocabulary mapped to index positions\n",
    "        self.word2index = {}\n",
    "        # TODO: populate self.word2index with indices for all the words in self.review_vocab\n",
    "        #       like you saw earlier in the notebook\n",
    "        for i,word in enumerate(self.review_vocab):\n",
    "            self.word2index[word] = i\n",
    "        \n",
    "        # Create a dictionary of labels mapped to index positions\n",
    "        self.label2index = {}\n",
    "        # TODO: do the same thing you did for self.word2index and self.review_vocab, \n",
    "        #       but for self.label2index and self.label_vocab instead\n",
    "        for i,word in enumerate(self.label_vocab):\n",
    "            self.label2index[word] = i\n",
    "         \n",
    "        \n",
    "    def init_network(self, input_nodes, hidden_nodes, output_nodes, learning_rate):\n",
    "        # Store the number of nodes in input, hidden, and output layers.\n",
    "        self.input_nodes = input_nodes\n",
    "        self.hidden_nodes = hidden_nodes\n",
    "        self.output_nodes = output_nodes\n",
    "\n",
    "        # Store the learning rate\n",
    "        self.learning_rate = learning_rate\n",
    "\n",
    "        # Initialize weights\n",
    "        \n",
    "        # TODO: initialize self.weights_0_1 as a matrix of zeros. These are the weights between\n",
    "        #       the input layer and the hidden layer.\n",
    "        self.weights_0_1 = np.zeros((self.input_nodes,self.hidden_nodes))\n",
    "        # TODO: initialize self.weights_1_2 as a matrix of random values. \n",
    "        #       These are the weights between the hidden layer and the output layer.\n",
    "        #self.weights_1_2 = np.random.rand(self.hidden_nodes, self.output_nodes)\n",
    "        \n",
    "        self.weights_1_2 = np.random.normal(0.0, self.output_nodes**-0.5, \n",
    "                                                (self.hidden_nodes, self.output_nodes))\n",
    "        \n",
    "        # TODO: Create the input layer, a two-dimensional matrix with shape \n",
    "        #       1 x input_nodes, with all values initialized to zero\n",
    "        self.layer_0 = np.zeros((1,input_nodes))\n",
    "    \n",
    "    #---------------------------------------------------------------------------  \n",
    "    def update_input_layer(self,review):\n",
    "        # TODO: You can copy most of the code you wrote for update_input_layer \n",
    "        #       earlier in this notebook. \n",
    "        # clear out previous state by resetting the layer to be all 0s\n",
    "        self.layer_0 *= 0\n",
    "    \n",
    "        # TODO: count how many times each word is used in the given review and store the results in layer_0 \n",
    "        for wd in review.split(' '):\n",
    "            if(wd in self.word2index.keys()):\n",
    "                self.layer_0[0][self.word2index[wd]] = 1\n",
    "            \n",
    "    #---------------------------------------------------------------------------\n",
    "    def get_target_for_label(self,label):\n",
    "        # TODO: Copy the code you wrote for get_target_for_label \n",
    "        #       earlier in this notebook. \n",
    "        if label == \"POSITIVE\":\n",
    "            return 1\n",
    "        else:\n",
    "            return 0\n",
    "        \n",
    "    #---------------------------------------------------------------------------\n",
    "    def sigmoid(self,x):\n",
    "        # TODO: Return the result of calculating the sigmoid activation function\n",
    "        #       shown in the lectures\n",
    "        return (1/(1+np.exp(-x)))\n",
    "    \n",
    "    def sigmoid_output_2_derivative(self,output):\n",
    "        # TODO: Return the derivative of the sigmoid activation function, \n",
    "        #       where \"output\" is the original output from the sigmoid function \n",
    "        return (output * (1-output) )\n",
    "\n",
    "    def train(self, training_reviews, training_labels):\n",
    "        \n",
    "        # make sure out we have a matching number of reviews and labels\n",
    "        assert(len(training_reviews) == len(training_labels))\n",
    "        \n",
    "        # Keep track of correct predictions to display accuracy during training \n",
    "        correct_so_far = 0\n",
    "        \n",
    "        # Remember when we started for printing time statistics\n",
    "        start = time.time()\n",
    "\n",
    "        # loop through all the given reviews and run a forward and backward pass,\n",
    "        # updating weights for every item\n",
    "        for i in range(len(training_reviews)):\n",
    "            \n",
    "            # TODO: Get the next review and its correct label\n",
    "            review = training_reviews[i]\n",
    "            label = training_labels[i]\n",
    "            \n",
    "            # TODO: Implement the forward pass through the network. \n",
    "            #       That means use the given review to update the input layer, \n",
    "            #       then calculate values for the hidden layer,\n",
    "            #       and finally calculate the output layer.\n",
    "            # \n",
    "            #       Do not use an activation function for the hidden layer,\n",
    "            #       but use the sigmoid activation function for the output layer.\n",
    "            self.update_input_layer(review)            \n",
    "            layer_1 = self.layer_0.dot(self.weights_0_1)\n",
    "            layer_2 = self.sigmoid(layer_1.dot(self.weights_1_2))\n",
    "                                    \n",
    "                                    \n",
    "            # TODO: Implement the back propagation pass here. \n",
    "            #       That means calculate the error for the forward pass's prediction\n",
    "            #       and update the weights in the network according to their\n",
    "            #       contributions toward the error, as calculated via the\n",
    "            #       gradient descent and back propagation algorithms you \n",
    "            #       learned in class.\n",
    "\n",
    "            layer_2_error = layer_2 - self.get_target_for_label(label)       # Output layer error == (True label - output)\n",
    "            layer_2_delta = layer_2_error * self.sigmoid_output_2_derivative(layer_2)\n",
    "\n",
    "            layer_1_error = layer_2_delta.dot(self.weights_1_2.T)\n",
    "            layer_1_delta = layer_1_error\n",
    "\n",
    "            self.weights_1_2 -= layer_1.T.dot(layer_2_delta) * self.learning_rate\n",
    "            self.weights_0_1 -= self.layer_0.T.dot(layer_1_delta) * self.learning_rate\n",
    "\n",
    "\n",
    "            \n",
    "            # TODO: Keep track of correct predictions. To determine if the prediction was\n",
    "            #       correct, check that the absolute value of the output error \n",
    "            #       is less than 0.5. If so, add one to the correct_so_far count.\n",
    "            if(layer_2 >= 0.5 and label == 'POSITIVE'):\n",
    "                correct_so_far += 1\n",
    "            elif(layer_2 < 0.5 and label == 'NEGATIVE'):\n",
    "                correct_so_far += 1            \n",
    "            \n",
    "            \n",
    "            # For debug purposes, print out our prediction accuracy and speed \n",
    "            # throughout the training process. \n",
    "\n",
    "            elapsed_time = float(time.time() - start)\n",
    "            reviews_per_second = i / elapsed_time if elapsed_time > 0 else 0\n",
    "            \n",
    "            sys.stdout.write(\"\\rProgress:\" + str(100 * i/float(len(training_reviews)))[:4] \\\n",
    "                             + \"% Speed(reviews/sec):\" + str(reviews_per_second)[0:5] \\\n",
    "                             + \" #Correct:\" + str(correct_so_far) + \" #Trained:\" + str(i+1) \\\n",
    "                             + \" Training Accuracy:\" + str(correct_so_far * 100 / float(i+1))[:4] + \"%\")\n",
    "            if(i % 2500 == 0):\n",
    "                print(\"\")\n",
    "    \n",
    "    def test(self, testing_reviews, testing_labels):\n",
    "        \"\"\"\n",
    "        Attempts to predict the labels for the given testing_reviews,\n",
    "        and uses the test_labels to calculate the accuracy of those predictions.\n",
    "        \"\"\"\n",
    "        \n",
    "        # keep track of how many correct predictions we make\n",
    "        correct = 0\n",
    "\n",
    "        # we'll time how many predictions per second we make\n",
    "        start = time.time()\n",
    "\n",
    "        # Loop through each of the given reviews and call run to predict\n",
    "        # its label. \n",
    "        for i in range(len(testing_reviews)):\n",
    "            pred = self.run(testing_reviews[i])\n",
    "            if(pred == testing_labels[i]):\n",
    "                correct += 1\n",
    "            \n",
    "            # For debug purposes, print out our prediction accuracy and speed \n",
    "            # throughout the prediction process. \n",
    "\n",
    "            elapsed_time = float(time.time() - start)\n",
    "            reviews_per_second = i / elapsed_time if elapsed_time > 0 else 0\n",
    "            \n",
    "            sys.stdout.write(\"\\rProgress:\" + str(100 * i/float(len(testing_reviews)))[:4] \\\n",
    "                             + \"% Speed(reviews/sec):\" + str(reviews_per_second)[0:5] \\\n",
    "                             + \" #Correct:\" + str(correct) + \" #Tested:\" + str(i+1) \\\n",
    "                             + \" Testing Accuracy:\" + str(correct * 100 / float(i+1))[:4] + \"%\")\n",
    "    \n",
    "    def run(self, review):\n",
    "        \"\"\"\n",
    "        Returns a POSITIVE or NEGATIVE prediction for the given review.\n",
    "        \"\"\"\n",
    "        # TODO: Run a forward pass through the network, like you did in the\n",
    "        #       \"train\" function. That means use the given review to \n",
    "        #       update the input layer, then calculate values for the hidden layer,\n",
    "        #       and finally calculate the output layer.\n",
    "        #\n",
    "        #       Note: The review passed into this function for prediction \n",
    "        #             might come from anywhere, so you should convert it \n",
    "        #             to lower case prior to using it.\n",
    "        \n",
    "        # TODO: The output layer should now contain a prediction. \n",
    "        #       Return `POSITIVE` for predictions greater-than-or-equal-to `0.5`, \n",
    "        #       and `NEGATIVE` otherwise.\n",
    "        \n",
    "        self.update_input_layer(review.lower())\n",
    "        layer_1 = self.layer_0.dot(self.weights_0_1)\n",
    "        layer_2 = self.sigmoid(layer_1.dot(self.weights_1_2))\n",
    "        \n",
    "        if(layer_2[0] >= 0.5):\n",
    "            return \"POSITIVE\"\n",
    "        else:\n",
    "            return \"NEGATIVE\"\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to recreate the network and train it. Notice we've gone back to the higher learning rate of `0.1`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:0.0% Speed(reviews/sec):0.0 #Correct:1 #Trained:1 Training Accuracy:100.%\n",
      "Progress:10.4% Speed(reviews/sec):86.56 #Correct:1808 #Trained:2501 Training Accuracy:72.2%\n",
      "Progress:20.8% Speed(reviews/sec):89.26 #Correct:3830 #Trained:5001 Training Accuracy:76.5%\n",
      "Progress:31.2% Speed(reviews/sec):90.46 #Correct:5912 #Trained:7501 Training Accuracy:78.8%\n",
      "Progress:41.6% Speed(reviews/sec):90.95 #Correct:8068 #Trained:10001 Training Accuracy:80.6%\n",
      "Progress:52.0% Speed(reviews/sec):91.18 #Correct:10203 #Trained:12501 Training Accuracy:81.6%\n",
      "Progress:62.5% Speed(reviews/sec):91.50 #Correct:12322 #Trained:15001 Training Accuracy:82.1%\n",
      "Progress:72.9% Speed(reviews/sec):91.54 #Correct:14443 #Trained:17501 Training Accuracy:82.5%\n",
      "Progress:83.3% Speed(reviews/sec):91.02 #Correct:16611 #Trained:20001 Training Accuracy:83.0%\n",
      "Progress:93.7% Speed(reviews/sec):89.71 #Correct:18788 #Trained:22501 Training Accuracy:83.4%\n",
      "Progress:99.9% Speed(reviews/sec):88.75 #Correct:20112 #Trained:24000 Training Accuracy:83.8%"
     ]
    }
   ],
   "source": [
    "mlp = SentimentNetwork(reviews[:-1000],labels[:-1000], learning_rate=0.1)\n",
    "mlp.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That should have trained much better than the earlier attempts. It's still not wonderful, but it should have improved dramatically. Run the following cell to test your model with 1000 predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:99.9% Speed(reviews/sec):587.5 #Correct:854 #Tested:1000 Testing Accuracy:85.4%"
     ]
    }
   ],
   "source": [
    "mlp.test(reviews[-1000:],labels[-1000:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# End of Project 4. \n",
    "## Andrew's solution was actually in the previous video, so rewatch that video if you had any problems with that project. Then continue on to the next lesson.\n",
    "# Analyzing Inefficiencies in our Network<a id='lesson_5'></a>\n",
    "The following cells include the code Andrew shows in the next video. We've included it here so you can run the cells along with the video without having to type in everything."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl4AAAEoCAYAAACJsv/HAAAABGdBTUEAALGPC/xhBQAAACBjSFJN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\nAghYEyB4WesY9SKAAAIIIICAWYFnu/zIiInRwogAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(filename='sentiment_network_sparse.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "layer_0 = np.zeros(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.])"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "layer_0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "layer_0[4] = 1\n",
    "layer_0[9] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.,  0.,  0.,  0.,  1.,  0.,  0.,  0.,  0.,  1.])"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "layer_0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "weights_0_1 = np.random.randn(10,5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.10503756,  0.44222989,  0.24392938, -0.55961832,  0.21389503])"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "layer_0.dot(weights_0_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "indices = [4,9]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "layer_1 = np.zeros(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for index in indices:\n",
    "    layer_1 += (1 * weights_0_1[index])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.10503756,  0.44222989,  0.24392938, -0.55961832,  0.21389503])"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "layer_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjsAAAEpCAYAAAB1IONWAAAABGdBTUEAALGPC/xhBQAAACBjSFJN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\nRK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(filename='sentiment_network_sparse_2.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "layer_1 = np.zeros(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for index in indices:\n",
    "    layer_1 += (weights_0_1[index])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.10503756,  0.44222989,  0.24392938, -0.55961832,  0.21389503])"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "layer_1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 5: Making our Network More Efficient<a id='project_5'></a>\n",
    "**TODO:** Make the `SentimentNetwork` class more efficient by eliminating unnecessary multiplications and additions that occur during forward and backward propagation. To do that, you can do the following:\n",
    "* Copy the `SentimentNetwork` class from the previous project into the following cell.\n",
    "* Remove the `update_input_layer` function - you will not need it in this version.\n",
    "* Modify `init_network`:\n",
    ">* You no longer need a separate input layer, so remove any mention of `self.layer_0`\n",
    ">* You will be dealing with the old hidden layer more directly, so create `self.layer_1`, a two-dimensional matrix with shape 1 x hidden_nodes, with all values initialized to zero\n",
    "* Modify `train`:\n",
    ">* Change the name of the input parameter `training_reviews` to `training_reviews_raw`. This will help with the next step.\n",
    ">* At the beginning of the function, you'll want to preprocess your reviews to convert them to a list of indices (from `word2index`) that are actually used in the review. This is equivalent to what you saw in the video when Andrew set specific indices to 1. Your code should create a local `list` variable named `training_reviews` that should contain a `list` for each review in `training_reviews_raw`. Those lists should contain the indices for words found in the review.\n",
    ">* Remove call to `update_input_layer`\n",
    ">* Use `self`'s  `layer_1` instead of a local `layer_1` object.\n",
    ">* In the forward pass, replace the code that updates `layer_1` with new logic that only adds the weights for the indices used in the review.\n",
    ">* When updating `weights_0_1`, only update the individual weights that were used in the forward pass.\n",
    "* Modify `run`:\n",
    ">* Remove call to `update_input_layer` \n",
    ">* Use `self`'s  `layer_1` instead of a local `layer_1` object.\n",
    ">* Much like you did in `train`, you will need to pre-process the `review` so you can work with word indices, then update `layer_1` by adding weights for the indices used in the review."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO: -Copy the SentimentNetwork class from Project 4 lesson\n",
    "#       -Modify it according to the above instructions \n",
    "import time\n",
    "import sys\n",
    "import numpy as np\n",
    "\n",
    "# Encapsulate our neural network in a class\n",
    "class SentimentNetwork:\n",
    "    def __init__(self, reviews,labels,hidden_nodes = 10, learning_rate = 0.1):\n",
    "        \"\"\"Create a SentimenNetwork with the given settings\n",
    "        Args:\n",
    "            reviews(list) - List of reviews used for training\n",
    "            labels(list) - List of POSITIVE/NEGATIVE labels associated with the given reviews\n",
    "            hidden_nodes(int) - Number of nodes to create in the hidden layer\n",
    "            learning_rate(float) - Learning rate to use while training\n",
    "        \n",
    "        \"\"\"\n",
    "        # Assign a seed to our random number generator to ensure we get\n",
    "        # reproducable results during development \n",
    "        np.random.seed(1)\n",
    "\n",
    "        # process the reviews and their associated labels so that everything\n",
    "        # is ready for training\n",
    "        self.pre_process_data(reviews, labels)\n",
    "        \n",
    "        # Build the network to have the number of hidden nodes and the learning rate that\n",
    "        # were passed into this initializer. Make the same number of input nodes as\n",
    "        # there are vocabulary words and create a single output node.\n",
    "        self.init_network(len(self.review_vocab),hidden_nodes, 1, learning_rate)\n",
    "\n",
    "    def pre_process_data(self, reviews, labels):\n",
    "        \n",
    "        # populate review_vocab with all of the words in the given reviews\n",
    "        review_vocab = set()\n",
    "        for review in reviews:\n",
    "            for word in review.split(\" \"):\n",
    "                review_vocab.add(word)\n",
    "\n",
    "        # Convert the vocabulary set to a list so we can access words via indices\n",
    "        self.review_vocab = list(review_vocab)\n",
    "        \n",
    "        # populate label_vocab with all of the words in the given labels.\n",
    "        label_vocab = set()\n",
    "        for label in labels:\n",
    "            label_vocab.add(label)\n",
    "        \n",
    "        # Convert the label vocabulary set to a list so we can access labels via indices\n",
    "        self.label_vocab = list(label_vocab)\n",
    "        \n",
    "        # Store the sizes of the review and label vocabularies.\n",
    "        self.review_vocab_size = len(self.review_vocab)\n",
    "        self.label_vocab_size = len(self.label_vocab)\n",
    "        \n",
    "        # Create a dictionary of words in the vocabulary mapped to index positions\n",
    "        self.word2index = {}\n",
    "        for i, word in enumerate(self.review_vocab):\n",
    "            self.word2index[word] = i\n",
    "        \n",
    "        # Create a dictionary of labels mapped to index positions\n",
    "        self.label2index = {}\n",
    "        for i, label in enumerate(self.label_vocab):\n",
    "            self.label2index[label] = i\n",
    "\n",
    "    def init_network(self, input_nodes, hidden_nodes, output_nodes, learning_rate):\n",
    "        # Set number of nodes in input, hidden and output layers.\n",
    "        self.input_nodes = input_nodes\n",
    "        self.hidden_nodes = hidden_nodes\n",
    "        self.output_nodes = output_nodes\n",
    "\n",
    "        # Store the learning rate\n",
    "        self.learning_rate = learning_rate\n",
    "\n",
    "        # Initialize weights\n",
    "\n",
    "        # These are the weights between the input layer and the hidden layer.\n",
    "        self.weights_0_1 = np.zeros((self.input_nodes,self.hidden_nodes))\n",
    "\n",
    "        # These are the weights between the hidden layer and the output layer.\n",
    "        self.weights_1_2 = np.random.normal(0.0, self.output_nodes**-0.5, \n",
    "                                                (self.hidden_nodes, self.output_nodes))\n",
    "        \n",
    "        ## New for Project 5: Removed self.layer_0; added self.layer_1\n",
    "        # The input layer, a two-dimensional matrix with shape 1 x hidden_nodes\n",
    "        self.layer_1 = np.zeros((1,hidden_nodes))\n",
    "    \n",
    "    ## New for Project 5: Removed update_input_layer function\n",
    "    \n",
    "    def get_target_for_label(self,label):\n",
    "        if(label == 'POSITIVE'):\n",
    "            return 1\n",
    "        else:\n",
    "            return 0\n",
    "        \n",
    "    def sigmoid(self,x):\n",
    "        return 1 / (1 + np.exp(-x))\n",
    "    \n",
    "    def sigmoid_output_2_derivative(self,output):\n",
    "        return output * (1 - output)\n",
    "    \n",
    "    ## New for Project 5: changed name of first parameter form 'training_reviews' \n",
    "    #                     to 'training_reviews_raw'\n",
    "    def train(self, training_reviews_raw, training_labels):\n",
    "\n",
    "        ## New for Project 5: pre-process training reviews so we can deal \n",
    "        #                     directly with the indices of non-zero inputs\n",
    "        training_reviews = list()\n",
    "        for review in training_reviews_raw:\n",
    "            indices = set()\n",
    "            for word in review.split(\" \"):\n",
    "                if(word in self.word2index.keys()):\n",
    "                    indices.add(self.word2index[word])\n",
    "            training_reviews.append(list(indices))\n",
    "\n",
    "        # make sure out we have a matching number of reviews and labels\n",
    "        assert(len(training_reviews) == len(training_labels))\n",
    "        \n",
    "        # Keep track of correct predictions to display accuracy during training \n",
    "        correct_so_far = 0\n",
    "\n",
    "        # Remember when we started for printing time statistics\n",
    "        start = time.time()\n",
    "        \n",
    "        # loop through all the given reviews and run a forward and backward pass,\n",
    "        # updating weights for every item\n",
    "        for i in range(len(training_reviews)):\n",
    "            \n",
    "            # Get the next review and its correct label\n",
    "            review = training_reviews[i]\n",
    "            label = training_labels[i]\n",
    "            \n",
    "            #### Implement the forward pass here ####\n",
    "            ### Forward pass ###\n",
    "\n",
    "            ## New for Project 5: Removed call to 'update_input_layer' function\n",
    "            #                     because 'layer_0' is no longer used\n",
    "\n",
    "            # Hidden layer\n",
    "            ## New for Project 5: Add in only the weights for non-zero items\n",
    "            self.layer_1 *= 0\n",
    "            for index in review:\n",
    "                self.layer_1 += self.weights_0_1[index]\n",
    "\n",
    "            # Output layer\n",
    "            ## New for Project 5: changed to use 'self.layer_1' instead of 'local layer_1'\n",
    "            layer_2 = self.sigmoid(self.layer_1.dot(self.weights_1_2))            \n",
    "            \n",
    "            #### Implement the backward pass here ####\n",
    "            ### Backward pass ###\n",
    "\n",
    "            # Output error\n",
    "            layer_2_error = layer_2 - self.get_target_for_label(label) # Output layer error is the difference between desired target and actual output.\n",
    "            layer_2_delta = layer_2_error * self.sigmoid_output_2_derivative(layer_2)\n",
    "\n",
    "            # Backpropagated error\n",
    "            layer_1_error = layer_2_delta.dot(self.weights_1_2.T) # errors propagated to the hidden layer\n",
    "            layer_1_delta = layer_1_error # hidden layer gradients - no nonlinearity so it's the same as the error\n",
    "\n",
    "            # Update the weights\n",
    "            ## New for Project 5: changed to use 'self.layer_1' instead of local 'layer_1'\n",
    "            self.weights_1_2 -= self.layer_1.T.dot(layer_2_delta) * self.learning_rate # update hidden-to-output weights with gradient descent step\n",
    "            \n",
    "            ## New for Project 5: Only update the weights that were used in the forward pass\n",
    "            for index in review:\n",
    "                self.weights_0_1[index] -= layer_1_delta[0] * self.learning_rate # update input-to-hidden weights with gradient descent step\n",
    "\n",
    "            # Keep track of correct predictions.\n",
    "            if(layer_2 >= 0.5 and label == 'POSITIVE'):\n",
    "                correct_so_far += 1\n",
    "            elif(layer_2 < 0.5 and label == 'NEGATIVE'):\n",
    "                correct_so_far += 1\n",
    "            \n",
    "            # For debug purposes, print out our prediction accuracy and speed \n",
    "            # throughout the training process. \n",
    "            elapsed_time = float(time.time() - start)\n",
    "            reviews_per_second = i / elapsed_time if elapsed_time > 0 else 0\n",
    "            \n",
    "            sys.stdout.write(\"\\rProgress:\" + str(100 * i/float(len(training_reviews)))[:4] \\\n",
    "                             + \"% Speed(reviews/sec):\" + str(reviews_per_second)[0:5] \\\n",
    "                             + \" #Correct:\" + str(correct_so_far) + \" #Trained:\" + str(i+1) \\\n",
    "                             + \" Training Accuracy:\" + str(correct_so_far * 100 / float(i+1))[:4] + \"%\")\n",
    "            if(i % 2500 == 0):\n",
    "                print(\"\")\n",
    "    \n",
    "    def test(self, testing_reviews, testing_labels):\n",
    "        \"\"\"\n",
    "        Attempts to predict the labels for the given testing_reviews,\n",
    "        and uses the test_labels to calculate the accuracy of those predictions.\n",
    "        \"\"\"\n",
    "        \n",
    "        # keep track of how many correct predictions we make\n",
    "        correct = 0\n",
    "\n",
    "        # we'll time how many predictions per second we make\n",
    "        start = time.time()\n",
    "\n",
    "        # Loop through each of the given reviews and call run to predict\n",
    "        # its label. \n",
    "        for i in range(len(testing_reviews)):\n",
    "            pred = self.run(testing_reviews[i])\n",
    "            if(pred == testing_labels[i]):\n",
    "                correct += 1\n",
    "            \n",
    "            # For debug purposes, print out our prediction accuracy and speed \n",
    "            # throughout the prediction process. \n",
    "\n",
    "            elapsed_time = float(time.time() - start)\n",
    "            reviews_per_second = i / elapsed_time if elapsed_time > 0 else 0\n",
    "            \n",
    "            sys.stdout.write(\"\\rProgress:\" + str(100 * i/float(len(testing_reviews)))[:4] \\\n",
    "                             + \"% Speed(reviews/sec):\" + str(reviews_per_second)[0:5] \\\n",
    "                             + \" #Correct:\" + str(correct) + \" #Tested:\" + str(i+1) \\\n",
    "                             + \" Testing Accuracy:\" + str(correct * 100 / float(i+1))[:4] + \"%\")\n",
    "    \n",
    "    def run(self, review):\n",
    "        \"\"\"\n",
    "        Returns a POSITIVE or NEGATIVE prediction for the given review.\n",
    "        \"\"\"\n",
    "        # Run a forward pass through the network, like in the \"train\" function.\n",
    "        \n",
    "        ## New for Project 5: Removed call to update_input_layer function\n",
    "        #                     because layer_0 is no longer used\n",
    "\n",
    "        # Hidden layer\n",
    "        ## New for Project 5: Identify the indices used in the review and then add\n",
    "        #                     just those weights to layer_1 \n",
    "        self.layer_1 *= 0\n",
    "        unique_indices = set()\n",
    "        for word in review.lower().split(\" \"):\n",
    "            if word in self.word2index.keys():\n",
    "                unique_indices.add(self.word2index[word])\n",
    "        for index in unique_indices:\n",
    "            self.layer_1 += self.weights_0_1[index]\n",
    "        \n",
    "        # Output layer\n",
    "        ## New for Project 5: changed to use self.layer_1 instead of local layer_1\n",
    "        layer_2 = self.sigmoid(self.layer_1.dot(self.weights_1_2))\n",
    "        \n",
    "        # Return POSITIVE for values above greater-than-or-equal-to 0.5 in the output layer;\n",
    "        # return NEGATIVE for other values\n",
    "        if(layer_2[0] >= 0.5):\n",
    "            return \"POSITIVE\"\n",
    "        else:\n",
    "            return \"NEGATIVE\"\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to recreate the network and train it once again."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:0.0% Speed(reviews/sec):0.0 #Correct:1 #Trained:1 Training Accuracy:100.%\n",
      "Progress:10.4% Speed(reviews/sec):848.1 #Correct:1824 #Trained:2501 Training Accuracy:72.9%\n",
      "Progress:20.8% Speed(reviews/sec):745.7 #Correct:3821 #Trained:5001 Training Accuracy:76.4%\n",
      "Progress:31.2% Speed(reviews/sec):738.5 #Correct:5913 #Trained:7501 Training Accuracy:78.8%\n",
      "Progress:41.6% Speed(reviews/sec):766.7 #Correct:8041 #Trained:10001 Training Accuracy:80.4%\n",
      "Progress:52.0% Speed(reviews/sec):748.6 #Correct:10165 #Trained:12501 Training Accuracy:81.3%\n",
      "Progress:62.5% Speed(reviews/sec):762.1 #Correct:12304 #Trained:15001 Training Accuracy:82.0%\n",
      "Progress:72.9% Speed(reviews/sec):772.9 #Correct:14429 #Trained:17501 Training Accuracy:82.4%\n",
      "Progress:83.3% Speed(reviews/sec):757.1 #Correct:16608 #Trained:20001 Training Accuracy:83.0%\n",
      "Progress:93.7% Speed(reviews/sec):763.9 #Correct:18802 #Trained:22501 Training Accuracy:83.5%\n",
      "Progress:99.9% Speed(reviews/sec):757.4 #Correct:20130 #Trained:24000 Training Accuracy:83.8%"
     ]
    }
   ],
   "source": [
    "mlp = SentimentNetwork(reviews[:-1000],labels[:-1000], learning_rate=0.1)\n",
    "mlp.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That should have trained much better than the earlier attempts. Run the following cell to test your model with 1000 predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress:99.9% Speed(reviews/sec):814.2 #Correct:852 #Tested:1000 Testing Accuracy:85.2%"
     ]
    }
   ],
   "source": [
    "mlp.test(reviews[-1000:],labels[-1000:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# End of Project 5. \n",
    "## Watch the next video to see Andrew's solution, then continue on to the next lesson.\n",
    "# Further Noise Reduction<a id='lesson_6'></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Image(filename='sentiment_network_sparse_2.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# words most frequently seen in a review with a \"POSITIVE\" label\n",
    "pos_neg_ratios.most_common()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# words most frequently seen in a review with a \"NEGATIVE\" label\n",
    "list(reversed(pos_neg_ratios.most_common()))[0:30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from bokeh.models import ColumnDataSource, LabelSet\n",
    "from bokeh.plotting import figure, show, output_file\n",
    "from bokeh.io import output_notebook\n",
    "output_notebook()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "hist, edges = np.histogram(list(map(lambda x:x[1],pos_neg_ratios.most_common())), density=True, bins=100, normed=True)\n",
    "\n",
    "p = figure(tools=\"pan,wheel_zoom,reset,save\",\n",
    "           toolbar_location=\"above\",\n",
    "           title=\"Word Positive/Negative Affinity Distribution\")\n",
    "p.quad(top=hist, bottom=0, left=edges[:-1], right=edges[1:], line_color=\"#555555\")\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "frequency_frequency = Counter()\n",
    "\n",
    "for word, cnt in total_counts.most_common():\n",
    "    frequency_frequency[cnt] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "hist, edges = np.histogram(list(map(lambda x:x[1],frequency_frequency.most_common())), density=True, bins=100, normed=True)\n",
    "\n",
    "p = figure(tools=\"pan,wheel_zoom,reset,save\",\n",
    "           toolbar_location=\"above\",\n",
    "           title=\"The frequency distribution of the words in our corpus\")\n",
    "p.quad(top=hist, bottom=0, left=edges[:-1], right=edges[1:], line_color=\"#555555\")\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 6: Reducing Noise by Strategically Reducing the Vocabulary<a id='project_6'></a>\n",
    "\n",
    "**TODO:** Improve `SentimentNetwork`'s performance by reducing more noise in the vocabulary. Specifically, do the following:\n",
    "* Copy the `SentimentNetwork` class from the previous project into the following cell.\n",
    "* Modify `pre_process_data`:\n",
    ">* Add two additional parameters: `min_count` and `polarity_cutoff`\n",
    ">* Calculate the positive-to-negative ratios of words used in the reviews. (You can use code you've written elsewhere in the notebook, but we are moving it into the class like we did with other helper code earlier.)\n",
    ">* Andrew's solution only calculates a postive-to-negative ratio for words that occur at least 50 times. This keeps the network from attributing too much sentiment to rarer words. You can choose to add this to your solution if you would like.  \n",
    ">* Change so words are only added to the vocabulary if they occur in the vocabulary more than `min_count` times.\n",
    ">* Change so words are only added to the vocabulary if the absolute value of their postive-to-negative ratio is at least `polarity_cutoff`\n",
    "* Modify `__init__`:\n",
    ">* Add the same two parameters (`min_count` and `polarity_cutoff`) and use them when you call `pre_process_data`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO: -Copy the SentimentNetwork class from Project 5 lesson\n",
    "#       -Modify it according to the above instructions "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to train your network with a small polarity cutoff."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mlp = SentimentNetwork(reviews[:-1000],labels[:-1000],min_count=20,polarity_cutoff=0.05,learning_rate=0.01)\n",
    "mlp.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And run the following cell to test it's performance. It should be "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mlp.test(reviews[-1000:],labels[-1000:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to train your network with a much larger polarity cutoff."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mlp = SentimentNetwork(reviews[:-1000],labels[:-1000],min_count=20,polarity_cutoff=0.8,learning_rate=0.01)\n",
    "mlp.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And run the following cell to test it's performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mlp.test(reviews[-1000:],labels[-1000:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# End of Project 6. \n",
    "## Watch the next video to see Andrew's solution, then continue on to the next lesson."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Analysis: What's Going on in the Weights?<a id='lesson_7'></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mlp_full = SentimentNetwork(reviews[:-1000],labels[:-1000],min_count=0,polarity_cutoff=0,learning_rate=0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mlp_full.train(reviews[:-1000],labels[:-1000])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Image(filename='sentiment_network_sparse.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_most_similar_words(focus = \"horrible\"):\n",
    "    most_similar = Counter()\n",
    "\n",
    "    for word in mlp_full.word2index.keys():\n",
    "        most_similar[word] = np.dot(mlp_full.weights_0_1[mlp_full.word2index[word]],mlp_full.weights_0_1[mlp_full.word2index[focus]])\n",
    "    \n",
    "    return most_similar.most_common()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "get_most_similar_words(\"excellent\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "get_most_similar_words(\"terrible\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.colors as colors\n",
    "\n",
    "words_to_visualize = list()\n",
    "for word, ratio in pos_neg_ratios.most_common(500):\n",
    "    if(word in mlp_full.word2index.keys()):\n",
    "        words_to_visualize.append(word)\n",
    "    \n",
    "for word, ratio in list(reversed(pos_neg_ratios.most_common()))[0:500]:\n",
    "    if(word in mlp_full.word2index.keys()):\n",
    "        words_to_visualize.append(word)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pos = 0\n",
    "neg = 0\n",
    "\n",
    "colors_list = list()\n",
    "vectors_list = list()\n",
    "for word in words_to_visualize:\n",
    "    if word in pos_neg_ratios.keys():\n",
    "        vectors_list.append(mlp_full.weights_0_1[mlp_full.word2index[word]])\n",
    "        if(pos_neg_ratios[word] > 0):\n",
    "            pos+=1\n",
    "            colors_list.append(\"#00ff00\")\n",
    "        else:\n",
    "            neg+=1\n",
    "            colors_list.append(\"#000000\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.manifold import TSNE\n",
    "tsne = TSNE(n_components=2, random_state=0)\n",
    "words_top_ted_tsne = tsne.fit_transform(vectors_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "p = figure(tools=\"pan,wheel_zoom,reset,save\",\n",
    "           toolbar_location=\"above\",\n",
    "           title=\"vector T-SNE for most polarized words\")\n",
    "\n",
    "source = ColumnDataSource(data=dict(x1=words_top_ted_tsne[:,0],\n",
    "                                    x2=words_top_ted_tsne[:,1],\n",
    "                                    names=words_to_visualize))\n",
    "\n",
    "p.scatter(x=\"x1\", y=\"x2\", size=8, source=source,color=colors_list)\n",
    "\n",
    "word_labels = LabelSet(x=\"x1\", y=\"x2\", text=\"names\", y_offset=6,\n",
    "                  text_font_size=\"8pt\", text_color=\"#555555\",\n",
    "                  source=source, text_align='center')\n",
    "p.add_layout(word_labels)\n",
    "\n",
    "show(p)\n",
    "\n",
    "# green indicates positive words, black indicates negative words"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}