sentement_analysis_keras.ipynb

{
  "cells": [
    {
      "metadata": {
        "collapsed": true
      },
      "cell_type": "markdown",
      "source": "# Sentement Analysis"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Data analysis"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from keras.datasets import imdb",
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Using TensorFlow backend.\n",
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "top_words = 1000\n(x_train,y_train),(x_test,y_test) = imdb.load_data(num_words=top_words)",
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Downloading data from https://s3.amazonaws.com/text-datasets/imdb.npz\n17465344/17464789 [==============================] - 3s 0us/step\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "print(x_train[0])",
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": "[1, 14, 22, 16, 43, 530, 973, 2, 2, 65, 458, 2, 66, 2, 4, 173, 36, 256, 5, 25, 100, 43, 838, 112, 50, 670, 2, 9, 35, 480, 284, 5, 150, 4, 172, 112, 167, 2, 336, 385, 39, 4, 172, 2, 2, 17, 546, 38, 13, 447, 4, 192, 50, 16, 6, 147, 2, 19, 14, 22, 4, 2, 2, 469, 4, 22, 71, 87, 12, 16, 43, 530, 38, 76, 15, 13, 2, 4, 22, 17, 515, 17, 12, 16, 626, 18, 2, 5, 62, 386, 12, 8, 316, 8, 106, 5, 4, 2, 2, 16, 480, 66, 2, 33, 4, 130, 12, 16, 38, 619, 5, 25, 124, 51, 36, 135, 48, 25, 2, 33, 6, 22, 12, 215, 28, 77, 52, 5, 14, 407, 16, 82, 2, 8, 4, 107, 117, 2, 15, 256, 4, 2, 7, 2, 5, 723, 36, 71, 43, 530, 476, 26, 400, 317, 46, 7, 4, 2, 2, 13, 104, 88, 4, 381, 15, 297, 98, 32, 2, 56, 26, 141, 6, 194, 2, 18, 4, 226, 22, 21, 134, 476, 26, 480, 5, 144, 30, 2, 18, 51, 36, 28, 224, 92, 25, 104, 4, 226, 65, 16, 38, 2, 88, 12, 16, 283, 5, 16, 2, 113, 103, 32, 15, 16, 2, 19, 178, 32]\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "print(y_train[0])",
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": "1\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "imdb.get_word_index()",
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Downloading data from https://s3.amazonaws.com/text-datasets/imdb_word_index.json\n1646592/1641221 [==============================] - 0s 0us/step\n",
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "execution_count": 11,
          "data": {
            "text/plain": "{'fawn': 34701,\n 'tsukino': 52006,\n 'nunnery': 52007,\n 'sonja': 16816,\n 'vani': 63951,\n 'woods': 1408,\n 'spiders': 16115,\n 'hanging': 2345,\n 'woody': 2289,\n 'trawling': 52008,\n \"hold's\": 52009,\n 'comically': 11307,\n 'localized': 40830,\n 'disobeying': 30568,\n \"'royale\": 52010,\n \"harpo's\": 40831,\n 'canet': 52011,\n 'aileen': 19313,\n 'acurately': 52012,\n \"diplomat's\": 52013,\n 'rickman': 25242,\n 'arranged': 6746,\n 'rumbustious': 52014,\n 'familiarness': 52015,\n \"spider'\": 52016,\n 'hahahah': 68804,\n \"wood'\": 52017,\n 'transvestism': 40833,\n \"hangin'\": 34702,\n 'bringing': 2338,\n 'seamier': 40834,\n 'wooded': 34703,\n 'bravora': 52018,\n 'grueling': 16817,\n 'wooden': 1636,\n 'wednesday': 16818,\n \"'prix\": 52019,\n 'altagracia': 34704,\n 'circuitry': 52020,\n 'crotch': 11585,\n 'busybody': 57766,\n \"tart'n'tangy\": 52021,\n 'burgade': 14129,\n 'thrace': 52023,\n \"tom's\": 11038,\n 'snuggles': 52025,\n 'francesco': 29114,\n 'complainers': 52027,\n 'templarios': 52125,\n '272': 40835,\n '273': 52028,\n 'zaniacs': 52130,\n '275': 34706,\n 'consenting': 27631,\n 'snuggled': 40836,\n 'inanimate': 15492,\n 'uality': 52030,\n 'bronte': 11926,\n 'errors': 4010,\n 'dialogs': 3230,\n \"yomada's\": 52031,\n \"madman's\": 34707,\n 'dialoge': 30585,\n 'usenet': 52033,\n 'videodrome': 40837,\n \"kid'\": 26338,\n 'pawed': 52034,\n \"'girlfriend'\": 30569,\n \"'pleasure\": 52035,\n \"'reloaded'\": 52036,\n \"kazakos'\": 40839,\n 'rocque': 52037,\n 'mailings': 52038,\n 'brainwashed': 11927,\n 'mcanally': 16819,\n \"tom''\": 52039,\n 'kurupt': 25243,\n 'affiliated': 21905,\n 'babaganoosh': 52040,\n \"noe's\": 40840,\n 'quart': 40841,\n 'kids': 359,\n 'uplifting': 5034,\n 'controversy': 7093,\n 'kida': 21906,\n 'kidd': 23379,\n \"error'\": 52041,\n 'neurologist': 52042,\n 'spotty': 18510,\n 'cobblers': 30570,\n 'projection': 9878,\n 'fastforwarding': 40842,\n 'sters': 52043,\n \"eggar's\": 52044,\n 'etherything': 52045,\n 'gateshead': 40843,\n 'airball': 34708,\n 'unsinkable': 25244,\n 'stern': 7180,\n \"cervi's\": 52046,\n 'dnd': 40844,\n 'dna': 11586,\n 'insecurity': 20598,\n \"'reboot'\": 52047,\n 'trelkovsky': 11037,\n 'jaekel': 52048,\n 'sidebars': 52049,\n \"sforza's\": 52050,\n 'distortions': 17633,\n 'mutinies': 52051,\n 'sermons': 30602,\n '7ft': 40846,\n 'boobage': 52052,\n \"o'bannon's\": 52053,\n 'populations': 23380,\n 'chulak': 52054,\n 'mesmerize': 27633,\n 'quinnell': 52055,\n 'yahoo': 10307,\n 'meteorologist': 52057,\n 'beswick': 42577,\n 'boorman': 15493,\n 'voicework': 40847,\n \"ster'\": 52058,\n 'blustering': 22922,\n 'hj': 52059,\n 'intake': 27634,\n 'morally': 5621,\n 'jumbling': 40849,\n 'bowersock': 52060,\n \"'porky's'\": 52061,\n 'gershon': 16821,\n 'ludicrosity': 40850,\n 'coprophilia': 52062,\n 'expressively': 40851,\n \"india's\": 19500,\n \"post's\": 34710,\n 'wana': 52063,\n 'wang': 5283,\n 'wand': 30571,\n 'wane': 25245,\n 'edgeways': 52321,\n 'titanium': 34711,\n 'pinta': 40852,\n 'want': 178,\n 'pinto': 30572,\n 'whoopdedoodles': 52065,\n 'tchaikovsky': 21908,\n 'travel': 2103,\n \"'victory'\": 52066,\n 'copious': 11928,\n 'gouge': 22433,\n \"chapters'\": 52067,\n 'barbra': 6702,\n 'uselessness': 30573,\n \"wan'\": 52068,\n 'assimilated': 27635,\n 'petiot': 16116,\n 'most\\x85and': 52069,\n 'dinosaurs': 3930,\n 'wrong': 352,\n 'seda': 52070,\n 'stollen': 52071,\n 'sentencing': 34712,\n 'ouroboros': 40853,\n 'assimilates': 40854,\n 'colorfully': 40855,\n 'glenne': 27636,\n 'dongen': 52072,\n 'subplots': 4760,\n 'kiloton': 52073,\n 'chandon': 23381,\n \"effect'\": 34713,\n 'snugly': 27637,\n 'kuei': 40856,\n 'welcomed': 9092,\n 'dishonor': 30071,\n 'concurrence': 52075,\n 'stoicism': 23382,\n \"guys'\": 14896,\n \"beroemd'\": 52077,\n 'butcher': 6703,\n \"melfi's\": 40857,\n 'aargh': 30623,\n 'playhouse': 20599,\n 'wickedly': 11308,\n 'fit': 1180,\n 'labratory': 52078,\n 'lifeline': 40859,\n 'screaming': 1927,\n 'fix': 4287,\n 'cineliterate': 52079,\n 'fic': 52080,\n 'fia': 52081,\n 'fig': 34714,\n 'fmvs': 52082,\n 'fie': 52083,\n 'reentered': 52084,\n 'fin': 30574,\n 'doctresses': 52085,\n 'fil': 52086,\n 'zucker': 12606,\n 'ached': 31931,\n 'counsil': 52088,\n 'paterfamilias': 52089,\n 'songwriter': 13885,\n 'shivam': 34715,\n 'hurting': 9654,\n 'effects': 299,\n 'slauther': 52090,\n \"'flame'\": 52091,\n 'sommerset': 52092,\n 'interwhined': 52093,\n 'whacking': 27638,\n 'bartok': 52094,\n 'barton': 8775,\n 'frewer': 21909,\n \"fi'\": 52095,\n 'ingrid': 6192,\n 'stribor': 30575,\n 'approporiately': 52096,\n 'wobblyhand': 52097,\n 'tantalisingly': 52098,\n 'ankylosaurus': 52099,\n 'parasites': 17634,\n 'childen': 52100,\n \"jenkins'\": 52101,\n 'metafiction': 52102,\n 'golem': 17635,\n 'indiscretion': 40860,\n \"reeves'\": 23383,\n \"inamorata's\": 57781,\n 'brittannica': 52104,\n 'adapt': 7916,\n \"russo's\": 30576,\n 'guitarists': 48246,\n 'abbott': 10553,\n 'abbots': 40861,\n 'lanisha': 17649,\n 'magickal': 40863,\n 'mattter': 52105,\n \"'willy\": 52106,\n 'pumpkins': 34716,\n 'stuntpeople': 52107,\n 'estimate': 30577,\n 'ugghhh': 40864,\n 'gameplay': 11309,\n \"wern't\": 52108,\n \"n'sync\": 40865,\n 'sickeningly': 16117,\n 'chiara': 40866,\n 'disturbed': 4011,\n 'portmanteau': 40867,\n 'ineffectively': 52109,\n \"duchonvey's\": 82143,\n \"nasty'\": 37519,\n 'purpose': 1285,\n 'lazers': 52112,\n 'lightened': 28105,\n 'kaliganj': 52113,\n 'popularism': 52114,\n \"damme's\": 18511,\n 'stylistics': 30578,\n 'mindgaming': 52115,\n 'spoilerish': 46449,\n \"'corny'\": 52117,\n 'boerner': 34718,\n 'olds': 6792,\n 'bakelite': 52118,\n 'renovated': 27639,\n 'forrester': 27640,\n \"lumiere's\": 52119,\n 'gaskets': 52024,\n 'needed': 884,\n 'smight': 34719,\n 'master': 1297,\n \"edie's\": 25905,\n 'seeber': 40868,\n 'hiya': 52120,\n 'fuzziness': 52121,\n 'genesis': 14897,\n 'rewards': 12607,\n 'enthrall': 30579,\n \"'about\": 40869,\n \"recollection's\": 52122,\n 'mutilated': 11039,\n 'fatherlands': 52123,\n \"fischer's\": 52124,\n 'positively': 5399,\n '270': 34705,\n 'ahmed': 34720,\n 'zatoichi': 9836,\n 'bannister': 13886,\n 'anniversaries': 52127,\n \"helm's\": 30580,\n \"'work'\": 52128,\n 'exclaimed': 34721,\n \"'unfunny'\": 52129,\n '274': 52029,\n 'feeling': 544,\n \"wanda's\": 52131,\n 'dolan': 33266,\n '278': 52133,\n 'peacoat': 52134,\n 'brawny': 40870,\n 'mishra': 40871,\n 'worlders': 40872,\n 'protags': 52135,\n 'skullcap': 52136,\n 'dastagir': 57596,\n 'affairs': 5622,\n 'wholesome': 7799,\n 'hymen': 52137,\n 'paramedics': 25246,\n 'unpersons': 52138,\n 'heavyarms': 52139,\n 'affaire': 52140,\n 'coulisses': 52141,\n 'hymer': 40873,\n 'kremlin': 52142,\n 'shipments': 30581,\n 'pixilated': 52143,\n \"'00s\": 30582,\n 'diminishing': 18512,\n 'cinematic': 1357,\n 'resonates': 14898,\n 'simplify': 40874,\n \"nature'\": 40875,\n 'temptresses': 40876,\n 'reverence': 16822,\n 'resonated': 19502,\n 'dailey': 34722,\n '2\\x85': 52144,\n 'treize': 27641,\n 'majo': 52145,\n 'kiya': 21910,\n 'woolnough': 52146,\n 'thanatos': 39797,\n 'sandoval': 35731,\n 'dorama': 40879,\n \"o'shaughnessy\": 52147,\n 'tech': 4988,\n 'fugitives': 32018,\n 'teck': 30583,\n \"'e'\": 76125,\n 'doesn’t': 40881,\n 'purged': 52149,\n 'saying': 657,\n \"martians'\": 41095,\n 'norliss': 23418,\n 'dickey': 27642,\n 'dicker': 52152,\n \"'sependipity\": 52153,\n 'padded': 8422,\n 'ordell': 57792,\n \"sturges'\": 40882,\n 'independentcritics': 52154,\n 'tempted': 5745,\n \"atkinson's\": 34724,\n 'hounded': 25247,\n 'apace': 52155,\n 'clicked': 15494,\n \"'humor'\": 30584,\n \"martino's\": 17177,\n \"'supporting\": 52156,\n 'warmongering': 52032,\n \"zemeckis's\": 34725,\n 'lube': 21911,\n 'shocky': 52157,\n 'plate': 7476,\n 'plata': 40883,\n 'sturgess': 40884,\n \"nerds'\": 40885,\n 'plato': 20600,\n 'plath': 34726,\n 'platt': 40886,\n 'mcnab': 52159,\n 'clumsiness': 27643,\n 'altogether': 3899,\n 'massacring': 42584,\n 'bicenntinial': 52160,\n 'skaal': 40887,\n 'droning': 14360,\n 'lds': 8776,\n 'jaguar': 21912,\n \"cale's\": 34727,\n 'nicely': 1777,\n 'mummy': 4588,\n \"lot's\": 18513,\n 'patch': 10086,\n 'kerkhof': 50202,\n \"leader's\": 52161,\n \"'movie\": 27644,\n 'uncomfirmed': 52162,\n 'heirloom': 40888,\n 'wrangle': 47360,\n 'emotion\\x85': 52163,\n \"'stargate'\": 52164,\n 'pinoy': 40889,\n 'conchatta': 40890,\n 'broeke': 41128,\n 'advisedly': 40891,\n \"barker's\": 17636,\n 'descours': 52166,\n 'lots': 772,\n 'lotr': 9259,\n 'irs': 9879,\n 'lott': 52167,\n 'xvi': 40892,\n 'irk': 34728,\n 'irl': 52168,\n 'ira': 6887,\n 'belzer': 21913,\n 'irc': 52169,\n 'ire': 27645,\n 'requisites': 40893,\n 'discipline': 7693,\n 'lyoko': 52961,\n 'extend': 11310,\n 'nature': 873,\n \"'dickie'\": 52170,\n 'optimist': 40894,\n 'lapping': 30586,\n 'superficial': 3900,\n 'vestment': 52171,\n 'extent': 2823,\n 'tendons': 52172,\n \"heller's\": 52173,\n 'quagmires': 52174,\n 'miyako': 52175,\n 'moocow': 20601,\n \"coles'\": 52176,\n 'lookit': 40895,\n 'ravenously': 52177,\n 'levitating': 40896,\n 'perfunctorily': 52178,\n 'lookin': 30587,\n \"lot'\": 40898,\n 'lookie': 52179,\n 'fearlessly': 34870,\n 'libyan': 52181,\n 'fondles': 40899,\n 'gopher': 35714,\n 'wearying': 40901,\n \"nz's\": 52182,\n 'minuses': 27646,\n 'puposelessly': 52183,\n 'shandling': 52184,\n 'decapitates': 31268,\n 'humming': 11929,\n \"'nother\": 40902,\n 'smackdown': 21914,\n 'underdone': 30588,\n 'frf': 40903,\n 'triviality': 52185,\n 'fro': 25248,\n 'bothers': 8777,\n \"'kensington\": 52186,\n 'much': 73,\n 'muco': 34730,\n 'wiseguy': 22615,\n \"richie's\": 27648,\n 'tonino': 40904,\n 'unleavened': 52187,\n 'fry': 11587,\n \"'tv'\": 40905,\n 'toning': 40906,\n 'obese': 14361,\n 'sensationalized': 30589,\n 'spiv': 40907,\n 'spit': 6259,\n 'arkin': 7364,\n 'charleton': 21915,\n 'jeon': 16823,\n 'boardroom': 21916,\n 'doubts': 4989,\n 'spin': 3084,\n 'hepo': 53083,\n 'wildcat': 27649,\n 'venoms': 10584,\n 'misconstrues': 52191,\n 'mesmerising': 18514,\n 'misconstrued': 40908,\n 'rescinds': 52192,\n 'prostrate': 52193,\n 'majid': 40909,\n 'climbed': 16479,\n 'canoeing': 34731,\n 'majin': 52195,\n 'animie': 57804,\n 'sylke': 40910,\n 'conditioned': 14899,\n 'waddell': 40911,\n '3\\x85': 52196,\n 'hyperdrive': 41188,\n 'conditioner': 34732,\n 'bricklayer': 53153,\n 'hong': 2576,\n 'memoriam': 52198,\n 'inventively': 30592,\n \"levant's\": 25249,\n 'portobello': 20638,\n 'remand': 52200,\n 'mummified': 19504,\n 'honk': 27650,\n 'spews': 19505,\n 'visitations': 40912,\n 'mummifies': 52201,\n 'cavanaugh': 25250,\n 'zeon': 23385,\n \"jungle's\": 40913,\n 'viertel': 34733,\n 'frenchmen': 27651,\n 'torpedoes': 52202,\n 'schlessinger': 52203,\n 'torpedoed': 34734,\n 'blister': 69876,\n 'cinefest': 52204,\n 'furlough': 34735,\n 'mainsequence': 52205,\n 'mentors': 40914,\n 'academic': 9094,\n 'stillness': 20602,\n 'academia': 40915,\n 'lonelier': 52206,\n 'nibby': 52207,\n \"losers'\": 52208,\n 'cineastes': 40916,\n 'corporate': 4449,\n 'massaging': 40917,\n 'bellow': 30593,\n 'absurdities': 19506,\n 'expetations': 53241,\n 'nyfiken': 40918,\n 'mehras': 75638,\n 'lasse': 52209,\n 'visability': 52210,\n 'militarily': 33946,\n \"elder'\": 52211,\n 'gainsbourg': 19023,\n 'hah': 20603,\n 'hai': 13420,\n 'haj': 34736,\n 'hak': 25251,\n 'hal': 4311,\n 'ham': 4892,\n 'duffer': 53259,\n 'haa': 52213,\n 'had': 66,\n 'advancement': 11930,\n 'hag': 16825,\n \"hand'\": 25252,\n 'hay': 13421,\n 'mcnamara': 20604,\n \"mozart's\": 52214,\n 'duffel': 30731,\n 'haq': 30594,\n 'har': 13887,\n 'has': 44,\n 'hat': 2401,\n 'hav': 40919,\n 'haw': 30595,\n 'figtings': 52215,\n 'elders': 15495,\n 'underpanted': 52216,\n 'pninson': 52217,\n 'unequivocally': 27652,\n \"barbara's\": 23673,\n \"bello'\": 52219,\n 'indicative': 12997,\n 'yawnfest': 40920,\n 'hexploitation': 52220,\n \"loder's\": 52221,\n 'sleuthing': 27653,\n \"justin's\": 32622,\n \"'ball\": 52222,\n \"'summer\": 52223,\n \"'demons'\": 34935,\n \"mormon's\": 52225,\n \"laughton's\": 34737,\n 'debell': 52226,\n 'shipyard': 39724,\n 'unabashedly': 30597,\n 'disks': 40401,\n 'crowd': 2290,\n 'crowe': 10087,\n \"vancouver's\": 56434,\n 'mosques': 34738,\n 'crown': 6627,\n 'culpas': 52227,\n 'crows': 27654,\n 'surrell': 53344,\n 'flowless': 52229,\n 'sheirk': 52230,\n \"'three\": 40923,\n \"peterson'\": 52231,\n 'ooverall': 52232,\n 'perchance': 40924,\n 'bottom': 1321,\n 'chabert': 53363,\n 'sneha': 52233,\n 'inhuman': 13888,\n 'ichii': 52234,\n 'ursla': 52235,\n 'completly': 30598,\n 'moviedom': 40925,\n 'raddick': 52236,\n 'brundage': 51995,\n 'brigades': 40926,\n 'starring': 1181,\n \"'goal'\": 52237,\n 'caskets': 52238,\n 'willcock': 52239,\n \"threesome's\": 52240,\n \"mosque'\": 52241,\n \"cover's\": 52242,\n 'spaceships': 17637,\n 'anomalous': 40927,\n 'ptsd': 27655,\n 'shirdan': 52243,\n 'obscenity': 21962,\n 'lemmings': 30599,\n 'duccio': 30600,\n \"levene's\": 52244,\n \"'gorby'\": 52245,\n \"teenager's\": 25255,\n 'marshall': 5340,\n 'honeymoon': 9095,\n 'shoots': 3231,\n 'despised': 12258,\n 'okabasho': 52246,\n 'fabric': 8289,\n 'cannavale': 18515,\n 'raped': 3537,\n \"tutt's\": 52247,\n 'grasping': 17638,\n 'despises': 18516,\n \"thief's\": 40928,\n 'rapes': 8926,\n 'raper': 52248,\n \"eyre'\": 27656,\n 'walchek': 52249,\n \"elmo's\": 23386,\n 'perfumes': 40929,\n 'spurting': 21918,\n \"exposition'\\x85\": 52250,\n 'denoting': 52251,\n 'thesaurus': 34740,\n \"shoot'\": 40930,\n 'bonejack': 49759,\n 'simpsonian': 52253,\n 'hebetude': 30601,\n \"hallow's\": 34741,\n 'desperation\\x85': 52254,\n 'incinerator': 34742,\n 'congratulations': 10308,\n 'humbled': 52255,\n \"else's\": 5924,\n 'trelkovski': 40845,\n \"rape'\": 52256,\n \"'chapters'\": 59386,\n '1600s': 52257,\n 'martian': 7253,\n 'nicest': 25256,\n 'eyred': 52259,\n 'passenger': 9457,\n 'disgrace': 6041,\n 'moderne': 52260,\n 'barrymore': 5120,\n 'yankovich': 52261,\n 'moderns': 40931,\n 'studliest': 52262,\n 'bedsheet': 52263,\n 'decapitation': 14900,\n 'slurring': 52264,\n \"'nunsploitation'\": 52265,\n \"'character'\": 34743,\n 'cambodia': 9880,\n 'rebelious': 52266,\n 'pasadena': 27657,\n 'crowne': 40932,\n \"'bedchamber\": 52267,\n 'conjectural': 52268,\n 'appologize': 52269,\n 'halfassing': 52270,\n 'paycheque': 57816,\n 'palms': 20606,\n \"'islands\": 52271,\n 'hawked': 40933,\n 'palme': 21919,\n 'conservatively': 40934,\n 'larp': 64007,\n 'palma': 5558,\n 'smelling': 21920,\n 'aragorn': 12998,\n 'hawker': 52272,\n 'hawkes': 52273,\n 'explosions': 3975,\n 'loren': 8059,\n \"pyle's\": 52274,\n 'shootout': 6704,\n \"mike's\": 18517,\n \"driscoll's\": 52275,\n 'cogsworth': 40935,\n \"britian's\": 52276,\n 'childs': 34744,\n \"portrait's\": 52277,\n 'chain': 3626,\n 'whoever': 2497,\n 'puttered': 52278,\n 'childe': 52279,\n 'maywether': 52280,\n 'chair': 3036,\n \"rance's\": 52281,\n 'machu': 34745,\n 'ballet': 4517,\n 'grapples': 34746,\n 'summerize': 76152,\n 'freelance': 30603,\n \"andrea's\": 52283,\n '\\x91very': 52284,\n 'coolidge': 45879,\n 'mache': 18518,\n 'balled': 52285,\n 'grappled': 40937,\n 'macha': 18519,\n 'underlining': 21921,\n 'macho': 5623,\n 'oversight': 19507,\n 'machi': 25257,\n 'verbally': 11311,\n 'tenacious': 21922,\n 'windshields': 40938,\n 'paychecks': 18557,\n 'jerk': 3396,\n \"good'\": 11931,\n 'prancer': 34748,\n 'prances': 21923,\n 'olympus': 52286,\n 'lark': 21924,\n 'embark': 10785,\n 'gloomy': 7365,\n 'jehaan': 52287,\n 'turaqui': 52288,\n \"child'\": 20607,\n 'locked': 2894,\n 'pranced': 52289,\n 'exact': 2588,\n 'unattuned': 52290,\n 'minute': 783,\n 'skewed': 16118,\n 'hodgins': 40940,\n 'skewer': 34749,\n 'think\\x85': 52291,\n 'rosenstein': 38765,\n 'helmit': 52292,\n 'wrestlemanias': 34750,\n 'hindered': 16826,\n \"martha's\": 30604,\n 'cheree': 52293,\n \"pluckin'\": 52294,\n 'ogles': 40941,\n 'heavyweight': 11932,\n 'aada': 82190,\n 'chopping': 11312,\n 'strongboy': 61534,\n 'hegemonic': 41342,\n 'adorns': 40942,\n 'xxth': 41346,\n 'nobuhiro': 34751,\n 'capitães': 52298,\n 'kavogianni': 52299,\n 'antwerp': 13422,\n 'celebrated': 6538,\n 'roarke': 52300,\n 'baggins': 40943,\n 'cheeseburgers': 31270,\n 'matras': 52301,\n \"nineties'\": 52302,\n \"'craig'\": 52303,\n 'celebrates': 12999,\n 'unintentionally': 3383,\n 'drafted': 14362,\n 'climby': 52304,\n '303': 52305,\n 'oldies': 18520,\n 'climbs': 9096,\n 'honour': 9655,\n 'plucking': 34752,\n '305': 30074,\n 'address': 5514,\n 'menjou': 40944,\n \"'freak'\": 42592,\n 'dwindling': 19508,\n 'benson': 9458,\n 'white’s': 52307,\n 'shamelessness': 40945,\n 'impacted': 21925,\n 'upatz': 52308,\n 'cusack': 3840,\n \"flavia's\": 37567,\n 'effette': 52309,\n 'influx': 34753,\n 'boooooooo': 52310,\n 'dimitrova': 52311,\n 'houseman': 13423,\n 'bigas': 25259,\n 'boylen': 52312,\n 'phillipenes': 52313,\n 'fakery': 40946,\n \"grandpa's\": 27658,\n 'darnell': 27659,\n 'undergone': 19509,\n 'handbags': 52315,\n 'perished': 21926,\n 'pooped': 37778,\n 'vigour': 27660,\n 'opposed': 3627,\n 'etude': 52316,\n \"caine's\": 11799,\n 'doozers': 52317,\n 'photojournals': 34754,\n 'perishes': 52318,\n 'constrains': 34755,\n 'migenes': 40948,\n 'consoled': 30605,\n 'alastair': 16827,\n 'wvs': 52319,\n 'ooooooh': 52320,\n 'approving': 34756,\n 'consoles': 40949,\n 'disparagement': 52064,\n 'futureistic': 52322,\n 'rebounding': 52323,\n \"'date\": 52324,\n 'gregoire': 52325,\n 'rutherford': 21927,\n 'americanised': 34757,\n 'novikov': 82196,\n 'following': 1042,\n 'munroe': 34758,\n \"morita'\": 52326,\n 'christenssen': 52327,\n 'oatmeal': 23106,\n 'fossey': 25260,\n 'livered': 40950,\n 'listens': 13000,\n \"'marci\": 76164,\n \"otis's\": 52330,\n 'thanking': 23387,\n 'maude': 16019,\n 'extensions': 34759,\n 'ameteurish': 52332,\n \"commender's\": 52333,\n 'agricultural': 27661,\n 'convincingly': 4518,\n 'fueled': 17639,\n 'mahattan': 54014,\n \"paris's\": 40952,\n 'vulkan': 52336,\n 'stapes': 52337,\n 'odysessy': 52338,\n 'harmon': 12259,\n 'surfing': 4252,\n 'halloran': 23494,\n 'unbelieveably': 49580,\n \"'offed'\": 52339,\n 'quadrant': 30607,\n 'inhabiting': 19510,\n 'nebbish': 34760,\n 'forebears': 40953,\n 'skirmish': 34761,\n 'ocassionally': 52340,\n \"'resist\": 52341,\n 'impactful': 21928,\n 'spicier': 52342,\n 'touristy': 40954,\n \"'football'\": 52343,\n 'webpage': 40955,\n 'exurbia': 52345,\n 'jucier': 52346,\n 'professors': 14901,\n 'structuring': 34762,\n 'jig': 30608,\n 'overlord': 40956,\n 'disconnect': 25261,\n 'sniffle': 82201,\n 'slimeball': 40957,\n 'jia': 40958,\n 'milked': 16828,\n 'banjoes': 40959,\n 'jim': 1237,\n 'workforces': 52348,\n 'jip': 52349,\n 'rotweiller': 52350,\n 'mundaneness': 34763,\n \"'ninja'\": 52351,\n \"dead'\": 11040,\n \"cipriani's\": 40960,\n 'modestly': 20608,\n \"professor'\": 52352,\n 'shacked': 40961,\n 'bashful': 34764,\n 'sorter': 23388,\n 'overpowering': 16120,\n 'workmanlike': 18521,\n 'henpecked': 27662,\n 'sorted': 18522,\n \"jōb's\": 52354,\n \"'always\": 52355,\n \"'baptists\": 34765,\n 'dreamcatchers': 52356,\n \"'silence'\": 52357,\n 'hickory': 21929,\n 'fun\\x97yet': 52358,\n 'breakumentary': 52359,\n 'didn': 15496,\n 'didi': 52360,\n 'pealing': 52361,\n 'dispite': 40962,\n \"italy's\": 25262,\n 'instability': 21930,\n 'quarter': 6539,\n 'quartet': 12608,\n 'padmé': 52362,\n \"'bleedmedry\": 52363,\n 'pahalniuk': 52364,\n 'honduras': 52365,\n 'bursting': 10786,\n \"pablo's\": 41465,\n 'irremediably': 52367,\n 'presages': 40963,\n 'bowlegged': 57832,\n 'dalip': 65183,\n 'entering': 6260,\n 'newsradio': 76172,\n 'presaged': 54150,\n \"giallo's\": 27663,\n 'bouyant': 40964,\n 'amerterish': 52368,\n 'rajni': 18523,\n 'leeves': 30610,\n 'macauley': 34767,\n 'seriously': 612,\n 'sugercoma': 52369,\n 'grimstead': 52370,\n \"'fairy'\": 52371,\n 'zenda': 30611,\n \"'twins'\": 52372,\n 'realisation': 17640,\n 'highsmith': 27664,\n 'raunchy': 7817,\n 'incentives': 40965,\n 'flatson': 52374,\n 'snooker': 35097,\n 'crazies': 16829,\n 'crazier': 14902,\n 'grandma': 7094,\n 'napunsaktha': 52375,\n 'workmanship': 30612,\n 'reisner': 52376,\n \"sanford's\": 61306,\n '\\x91doña': 52377,\n 'modest': 6108,\n \"everything's\": 19153,\n 'hamer': 40966,\n \"couldn't'\": 52379,\n 'quibble': 13001,\n 'socking': 52380,\n 'tingler': 21931,\n 'gutman': 52381,\n 'lachlan': 40967,\n 'tableaus': 52382,\n 'headbanger': 52383,\n 'spoken': 2847,\n 'cerebrally': 34768,\n \"'road\": 23490,\n 'tableaux': 21932,\n \"proust's\": 40968,\n 'periodical': 40969,\n \"shoveller's\": 52385,\n 'tamara': 25263,\n 'affords': 17641,\n 'concert': 3249,\n \"yara's\": 87955,\n 'someome': 52386,\n 'lingering': 8424,\n \"abraham's\": 41511,\n 'beesley': 34769,\n 'cherbourg': 34770,\n 'kagan': 28624,\n 'snatch': 9097,\n \"miyazaki's\": 9260,\n 'absorbs': 25264,\n \"koltai's\": 40970,\n 'tingled': 64027,\n 'crossroads': 19511,\n 'rehab': 16121,\n 'falworth': 52389,\n 'sequals': 52390,\n ...}"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "word_dict = imdb.get_word_index()\nword_dict = { key:(value + 3) for key, value in word_dict.items() }\nword_dict[''] = 0  # Padding\nword_dict['>'] = 1 # Start\nword_dict['?'] = 2 # Unknown word\nreverse_word_dict = { value:key for key, value in word_dict.items() }\nprint(' '.join(reverse_word_dict[id] for id in x_train[0]))",
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": "> this film was just brilliant casting ? ? story direction ? really ? the part they played and you could just imagine being there robert ? is an amazing actor and now the same being director ? father came from the same ? ? as myself so i loved the fact there was a real ? with this film the ? ? throughout the film were great it was just brilliant so much that i ? the film as soon as it was released for ? and would recommend it to everyone to watch and the ? ? was amazing really ? at the end it was so sad and you know what they say if you ? at a film it must have been good and this definitely was also ? to the two little ? that played the ? of ? and paul they were just brilliant children are often left out of the ? ? i think because the stars that play them all ? up are such a big ? for the whole film but these children are amazing and should be ? for what they have done don't you think the whole story was so ? because it was true and was ? life after all that was ? with us all\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Data Preprocessing"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from keras.preprocessing import sequence\nmax_review_len = 500\nx_train = sequence.pad_sequences(x_train, maxlen=max_review_len)\nx_test = sequence.pad_sequences(x_test, maxlen=max_review_len)",
      "execution_count": 14,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Model creation and Training"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from keras.models import Sequential\nfrom keras.layers import Dense\nfrom keras.layers.embeddings import Embedding\nfrom keras.layers import Flatten",
      "execution_count": 15,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "embedding_vec_len=32\nmodel = Sequential()\n\nmodel.add(Embedding(top_words,embedding_vec_len,input_length=max_review_len))\nmodel.add(Flatten())\nmodel.add(Dense(16, activation='relu'))\nmodel.add(Dense(16, activation='relu'))\nmodel.add(Dense(1, activation='sigmoid'))\n\nmodel.compile('adam',loss='binary_crossentropy', metrics=['accuracy'])\n\nmodel.summary()",
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": "_________________________________________________________________\nLayer (type)                 Output Shape              Param #   \n=================================================================\nembedding_6 (Embedding)      (None, 500, 32)           32000     \n_________________________________________________________________\nflatten_6 (Flatten)          (None, 16000)             0         \n_________________________________________________________________\ndense_16 (Dense)             (None, 16)                256016    \n_________________________________________________________________\ndense_17 (Dense)             (None, 16)                272       \n_________________________________________________________________\ndense_18 (Dense)             (None, 1)                 17        \n=================================================================\nTotal params: 288,305\nTrainable params: 288,305\nNon-trainable params: 0\n_________________________________________________________________\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "hist = model.fit(x=x_train,y=y_train,validation_data=(x_test,y_test),epochs=2,batch_size=100)",
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Train on 25000 samples, validate on 25000 samples\nEpoch 1/2\n25000/25000 [==============================] - 9s 340us/step - loss: 0.5236 - acc: 0.7053 - val_loss: 0.3315 - val_acc: 0.8580\nEpoch 2/2\n25000/25000 [==============================] - 8s 337us/step - loss: 0.2833 - acc: 0.8822 - val_loss: 0.3279 - val_acc: 0.8578\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "import seaborn as sns\nimport matplotlib.pyplot as plt\n%matplotlib inline\n\nsns.set()\nacc = hist.history['acc']\nval = hist.history['val_acc']\nepochs = range(1, len(acc) + 1)\n\nplt.plot(epochs, acc, '-', label='Training accuracy')\nplt.plot(epochs, val, ':', label='Validation accuracy')\nplt.title('Training and Validation Accuracy')\nplt.xlabel('Epoch')\nplt.ylabel('Accuracy')\nplt.legend(loc='upper left')\nplt.plot()",
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 33,
          "data": {
            "text/plain": "[]"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "sns.set()\nacc = hist.history['loss']\nval = hist.history['val_loss']\nepochs = range(1, len(acc) + 1)\n\nplt.plot(epochs, acc, '-', label='Training loss')\nplt.plot(epochs, val, ':', label='Validation loss')\nplt.title('Training and Validation Loss')\nplt.xlabel('Epoch')\nplt.ylabel('Loss')\nplt.legend(loc='upper left')\nplt.plot()",
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 34,
          "data": {
            "text/plain": "[]"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "scores = model.evaluate(x_test, y_test, verbose=0)\nprint(\"Accuracy: %.2f%%\" % (scores[1] * 100))",
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Accuracy: 85.78%\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Test Case"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "import string\nimport numpy as np\n\ndef analyze(text):\n    # Prepare the input by removing punctuation characters, converting\n    # characters to lower case, and removing words containing numbers\n    translator = str.maketrans('', '', string.punctuation)\n    text = text.translate(translator)\n    text = text.lower().split(' ')\n    text = [word for word in text if word.isalpha()]\n\n    # Generate an input tensor\n    input = [1]\n    for word in text:\n        if word in word_dict and word_dict[word] < top_words:\n            input.append(word_dict[word])\n        else:\n            input.append(2)\n    padded_input = sequence.pad_sequences([input], maxlen=max_review_len)\n\n    # Invoke the model and return the result\n    result = model.predict(np.array([padded_input][0]))[0][0]\n    return result",
      "execution_count": 38,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "analyze('Azure notebooks are awesome.')",
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 39,
          "data": {
            "text/plain": "0.7455242"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "name": "python36",
      "display_name": "Python 3.6",
      "language": "python"
    },
    "language_info": {
      "mimetype": "text/x-python",
      "nbconvert_exporter": "python",
      "name": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.6",
      "file_extension": ".py",
      "codemirror_mode": {
        "version": 3,
        "name": "ipython"
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}