LDA_CustomFunctions.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IkrzJPAIKh_Q"
   },
   "source": [
    "## Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "9XbLh4z2Kcmy"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "plt.style.use(\"ggplot\")\n",
    "plt.usetex = True\n",
    "from tqdm.notebook import tqdm\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "V3fEriKCLDOD"
   },
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_20newsgroups\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "P_PTejQ8LEv0"
   },
   "outputs": [],
   "source": [
    "data, target = fetch_20newsgroups(shuffle=True, random_state=1,\n",
    "                             remove=('headers', 'footers', 'quotes'),\n",
    "                             return_X_y=True,\n",
    "                            )\n",
    "# data_samples = data[:n_samples]\n",
    "data_samples = data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "0AbsfmfMLGGC"
   },
   "outputs": [],
   "source": [
    "tf_vectorizer = CountVectorizer(max_df=0.95, min_df=2,\n",
    "                                max_features=10000,\n",
    "                                stop_words='english')\n",
    "tf = tf_vectorizer.fit_transform(data_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "hpsqtLH_LHZg"
   },
   "outputs": [],
   "source": [
    "vocabulary = tf_vectorizer.vocabulary_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OXBoFu57LIus",
    "outputId": "80affdc9-664f-4918-cf1e-fd8a242de42f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'sure': 8719,\n",
       " 'story': 8583,\n",
       " 'did': 3039,\n",
       " 'biased': 1580,\n",
       " 'disagree': 3087,\n",
       " 'statement': 8519,\n",
       " 'media': 5795,\n",
       " 'reputation': 7609,\n",
       " 'pro': 7078,\n",
       " 'israeli': 5000,\n",
       " 'world': 9794,\n",
       " 'having': 4386,\n",
       " 'lived': 5490,\n",
       " 'europe': 3538,\n",
       " 'realize': 7426,\n",
       " 'described': 2960,\n",
       " 'letter': 5398,\n",
       " 'occured': 6360,\n",
       " 'try': 9186,\n",
       " 'ignore': 4649,\n",
       " 'existance': 3603,\n",
       " 'europeans': 3540,\n",
       " 'degree': 2901,\n",
       " 'think': 8965,\n",
       " 'reason': 7430,\n",
       " 'report': 7589,\n",
       " 'clearly': 2191,\n",
       " 'atrocities': 1327,\n",
       " 'shame': 8149,\n",
       " 'austria': 1361,\n",
       " 'daily': 2775,\n",
       " 'reports': 7594,\n",
       " 'acts': 880,\n",
       " 'soldiers': 8353,\n",
       " 'received': 7444,\n",
       " 'government': 4210,\n",
       " 'makes': 5649,\n",
       " 'holocaust': 4507,\n",
       " 'guilt': 4296,\n",
       " 'away': 1399,\n",
       " 'look': 5532,\n",
       " 'jews': 5058,\n",
       " 'treating': 9144,\n",
       " 'races': 7343,\n",
       " 'got': 4205,\n",
       " 'power': 6964,\n",
       " 'unfortunate': 9304,\n",
       " 'yeah': 9930,\n",
       " 'expect': 3614,\n",
       " 'people': 6701,\n",
       " 'read': 7412,\n",
       " 'faq': 3728,\n",
       " 'actually': 882,\n",
       " 'accept': 827,\n",
       " 'hard': 4361,\n",
       " 'atheism': 1313,\n",
       " 'need': 6162,\n",
       " 'little': 5487,\n",
       " 'leap': 5361,\n",
       " 'faith': 3708,\n",
       " 'logic': 5521,\n",
       " 'runs': 7858,\n",
       " 'steam': 8544,\n",
       " 'jim': 5061,\n",
       " 'sorry': 8373,\n",
       " 'pity': 6814,\n",
       " 'feelings': 3773,\n",
       " 'denial': 2933,\n",
       " 'oh': 6388,\n",
       " 'just': 5132,\n",
       " 'pretend': 7034,\n",
       " 'end': 3430,\n",
       " 'happily': 4357,\n",
       " 'maybe': 5752,\n",
       " 'start': 8508,\n",
       " 'new': 6198,\n",
       " 'newsgroup': 6205,\n",
       " 'alt': 1043,\n",
       " 'atheist': 1314,\n",
       " 'won': 9773,\n",
       " 'bye': 1847,\n",
       " 'big': 1586,\n",
       " 'don': 3191,\n",
       " 'forget': 3931,\n",
       " 'iii': 4656,\n",
       " 'principle': 7060,\n",
       " 'strongest': 8612,\n",
       " 'points': 6887,\n",
       " 'like': 5446,\n",
       " 'know': 5235,\n",
       " 'ask': 1258,\n",
       " 'question': 7305,\n",
       " 'sort': 8374,\n",
       " 'arab': 1186,\n",
       " 'countries': 2623,\n",
       " 'want': 9620,\n",
       " 'continue': 2508,\n",
       " 'tank': 8822,\n",
       " 'israel': 4999,\n",
       " 'stop': 8573,\n",
       " 'asking': 1260,\n",
       " 'questions': 7309,\n",
       " 'work': 9785,\n",
       " 'treatment': 9145,\n",
       " 'decades': 2841,\n",
       " 'bad': 1442,\n",
       " 'begin': 1530,\n",
       " 'attack': 1331,\n",
       " 'group': 4271,\n",
       " 'stupid': 8631,\n",
       " 'center': 2010,\n",
       " 'policy': 6894,\n",
       " 'research': 7622,\n",
       " 'fancy': 3724,\n",
       " 'legitimate': 5382,\n",
       " 'proposal': 7163,\n",
       " 'change': 2049,\n",
       " 'att': 1328,\n",
       " 'product': 7104,\n",
       " 'area': 1201,\n",
       " 'priced': 7046,\n",
       " '1000': 35,\n",
       " 'suspect': 8737,\n",
       " 'clipper': 2204,\n",
       " 'phones': 6765,\n",
       " 'came': 1896,\n",
       " 'customer': 2737,\n",
       " 'key': 5179,\n",
       " 'automatically': 1373,\n",
       " 'authorities': 1364,\n",
       " 'aside': 1257,\n",
       " 'attempting': 1338,\n",
       " 'fed': 3763,\n",
       " 'direction': 3075,\n",
       " 'yes': 9938,\n",
       " 'technology': 8866,\n",
       " 'eventually': 3550,\n",
       " 'drive': 3236,\n",
       " 'cost': 2611,\n",
       " 'promote': 7142,\n",
       " 'widespread': 9708,\n",
       " 'use': 9367,\n",
       " 'present': 7017,\n",
       " 'man': 5654,\n",
       " 'street': 8596,\n",
       " 'going': 4192,\n",
       " 'purchase': 7233,\n",
       " 'crypto': 2700,\n",
       " 'telephone': 8879,\n",
       " 'especially': 3513,\n",
       " 'guy': 4303,\n",
       " 'probably': 7081,\n",
       " 'doesn': 3180,\n",
       " 'missing': 5928,\n",
       " 'real': 7423,\n",
       " 'gov': 4208,\n",
       " 'year': 9931,\n",
       " 'air': 995,\n",
       " 'tight': 9006,\n",
       " 'voice': 9547,\n",
       " 'privacy': 7074,\n",
       " 'phone': 6764,\n",
       " 'line': 5463,\n",
       " 'close': 2208,\n",
       " 'nearest': 6154,\n",
       " 'pc': 6675,\n",
       " 'scenario': 7957,\n",
       " 'extent': 3662,\n",
       " 'usage': 9365,\n",
       " 'underground': 9295,\n",
       " 'pgp': 6746,\n",
       " 'scoring': 7985,\n",
       " 'playoff': 6852,\n",
       " 'pool': 6907,\n",
       " 'unfortunately': 9305,\n",
       " 'time': 9010,\n",
       " 'right': 7733,\n",
       " 'certainly': 2022,\n",
       " 'post': 6943,\n",
       " 'rules': 7850,\n",
       " 'tomorrow': 9047,\n",
       " 'does': 3179,\n",
       " 'matter': 5744,\n",
       " 'll': 5498,\n",
       " 'enter': 3459,\n",
       " 'good': 4199,\n",
       " 'keith': 5165,\n",
       " 'keller': 5167,\n",
       " 'let': 5395,\n",
       " 'rangers': 7373,\n",
       " 'quakers': 7288,\n",
       " 'kkeller': 5223,\n",
       " 'mail': 5627,\n",
       " 'sas': 7923,\n",
       " 'upenn': 9345,\n",
       " 'edu': 3336,\n",
       " 'ivy': 5017,\n",
       " 'league': 5357,\n",
       " 'champs': 2045,\n",
       " 'smith': 8320,\n",
       " 'old': 6402,\n",
       " 'wasn': 9643,\n",
       " 'dead': 2820,\n",
       " 'tomb': 9045,\n",
       " 'magi': 5622,\n",
       " 'cult': 2722,\n",
       " 'jesus': 5053,\n",
       " 'appears': 1154,\n",
       " 'day': 2814,\n",
       " 'stay': 8533,\n",
       " 'normal': 6257,\n",
       " 'wonder': 9774,\n",
       " 'ok': 6396,\n",
       " 'record': 7459,\n",
       " 'shows': 8210,\n",
       " 'iisi': 4657,\n",
       " 'cache': 1869,\n",
       " 'small': 8312,\n",
       " 'attach': 1329,\n",
       " 'measured': 5783,\n",
       " 'programs': 7122,\n",
       " '64': 595,\n",
       " 'kb': 5159,\n",
       " 'varies': 9443,\n",
       " 'lot': 5550,\n",
       " 'app': 1145,\n",
       " 'ranging': 7374,\n",
       " '40': 475,\n",
       " 'average': 1386,\n",
       " '20': 257,\n",
       " '25': 315,\n",
       " 'subjective': 8640,\n",
       " 'difference': 3049,\n",
       " 'great': 4247,\n",
       " 'simple': 8245,\n",
       " 'card': 1935,\n",
       " 'transform': 9117,\n",
       " 'better': 1571,\n",
       " 'fpu': 3969,\n",
       " 'conventional': 2542,\n",
       " 'wisdom': 9751,\n",
       " 'says': 7945,\n",
       " 'cards': 1938,\n",
       " 'makers': 5648,\n",
       " 'offer': 6376,\n",
       " '32k': 409,\n",
       " '64k': 597,\n",
       " 'caches': 1870,\n",
       " 'bought': 1706,\n",
       " 'wave': 9655,\n",
       " '150': 112,\n",
       " 'absolutely': 809,\n",
       " 'problems': 7086,\n",
       " 'complete': 2355,\n",
       " 'check': 2080,\n",
       " 'called': 1889,\n",
       " 'performance': 6711,\n",
       " 'rating': 7389,\n",
       " 'numbers': 6296,\n",
       " 'cheers': 2085,\n",
       " 'file': 3809,\n",
       " 'converted': 2549,\n",
       " 'q0': 7255,\n",
       " 'uu': 9391,\n",
       " 'ut': 9382,\n",
       " 'ud': 9256,\n",
       " 'b8': 1423,\n",
       " 'zv': 9997,\n",
       " 'ik': 4659,\n",
       " 'q6': 7263,\n",
       " 'qh': 7272,\n",
       " 'lr': 5567,\n",
       " 'pa': 6549,\n",
       " 'ha': 4319,\n",
       " 'vr': 9567,\n",
       " 'vz': 9577,\n",
       " 'vf': 9485,\n",
       " 'fj': 3856,\n",
       " 'ei': 3362,\n",
       " 'yr': 9952,\n",
       " 'mr': 6055,\n",
       " 'mc': 5760,\n",
       " 'da': 2773,\n",
       " 'ti': 8998,\n",
       " '6v': 634,\n",
       " 'ej': 3367,\n",
       " 'gs': 4278,\n",
       " '2p': 373,\n",
       " 'vi': 9492,\n",
       " 'tr': 9086,\n",
       " 'cr': 2649,\n",
       " 'cc': 1988,\n",
       " 'xl': 9873,\n",
       " 'lj1': 5495,\n",
       " 'rb': 7398,\n",
       " 'cp': 2644,\n",
       " '5p': 570,\n",
       " 'dm': 3159,\n",
       " 'b9': 1427,\n",
       " 'f2': 3679,\n",
       " 'aj': 1000,\n",
       " 'hj': 4489,\n",
       " 'nr': 6281,\n",
       " 'hk': 4490,\n",
       " 'a2': 779,\n",
       " 'bz': 1850,\n",
       " 'gl': 4163,\n",
       " 'ge': 4092,\n",
       " 'gh': 4139,\n",
       " 'vp': 9566,\n",
       " 'mp': 6050,\n",
       " 'eb': 3314,\n",
       " 'pi': 6780,\n",
       " 'pq': 6973,\n",
       " 'uc': 9250,\n",
       " '3c': 445,\n",
       " '33': 410,\n",
       " 'ap': 1135,\n",
       " 'sounds': 8383,\n",
       " 'guessing': 4290,\n",
       " 'mean': 5774,\n",
       " 'peace': 6683,\n",
       " 'process': 7094,\n",
       " 'say': 7943,\n",
       " 'palestineans': 6574,\n",
       " 'founded': 3962,\n",
       " 'predictions': 6997,\n",
       " 'test': 8921,\n",
       " 'table': 8802,\n",
       " '18': 150,\n",
       " 'months': 6001,\n",
       " 'prove': 7187,\n",
       " 'fault': 3744,\n",
       " 'known': 5239,\n",
       " 'accepted': 830,\n",
       " 'terms': 8909,\n",
       " 'israelis': 5001,\n",
       " 'ready': 7421,\n",
       " 'politics': 6901,\n",
       " 'leadership': 5352,\n",
       " 'saying': 7944,\n",
       " 'reasons': 7435,\n",
       " 'make': 5645,\n",
       " 'arabs': 1189,\n",
       " 'negotiations': 6173,\n",
       " 'california': 1886,\n",
       " 'regards': 7498,\n",
       " 'shouldn': 8204,\n",
       " 'allowed': 1037,\n",
       " 'msg': 6060,\n",
       " 'force': 3922,\n",
       " 'food': 3915,\n",
       " 'enhance': 3452,\n",
       " 'salt': 7905,\n",
       " 'option': 6453,\n",
       " 'leave': 5366,\n",
       " 'include': 4731,\n",
       " 'packet': 6558,\n",
       " 'packets': 6559,\n",
       " 'desire': 2977,\n",
       " 'options': 6455,\n",
       " 'treat': 9142,\n",
       " 'way': 9659,\n",
       " 'wouldn': 9806,\n",
       " 'throat': 8988,\n",
       " 'wondering': 9777,\n",
       " 'shed': 8167,\n",
       " 'light': 5440,\n",
       " 'electronic': 3379,\n",
       " 'remember': 7555,\n",
       " 'total': 9072,\n",
       " 'mileage': 5883,\n",
       " 'kind': 5206,\n",
       " 'memory': 5816,\n",
       " 'stable': 8485,\n",
       " 'reliable': 7537,\n",
       " 'non': 6248,\n",
       " 'independent': 4752,\n",
       " 'outside': 6513,\n",
       " 'battery': 1499,\n",
       " '10': 33,\n",
       " 'years': 9932,\n",
       " 'life': 5434,\n",
       " 'vehicle': 9459,\n",
       " 'amazed': 1055,\n",
       " 'expected': 3616,\n",
       " 'length': 5386,\n",
       " 'doing': 3183,\n",
       " 'breaking': 1744,\n",
       " 'legal': 5377,\n",
       " 'selling': 8067,\n",
       " 'used': 9368,\n",
       " 'car': 1931,\n",
       " 'replaced': 7582,\n",
       " 'odometer': 6371,\n",
       " 'starts': 8513,\n",
       " 'miles': 5884,\n",
       " '100': 34,\n",
       " '200': 258,\n",
       " 'actual': 881,\n",
       " 'looks': 5535,\n",
       " 'fraud': 3983,\n",
       " 'fairly': 3707,\n",
       " 'easy': 3309,\n",
       " 'price': 7045,\n",
       " 'tell': 8883,\n",
       " 'buyer': 1839,\n",
       " 'thanks': 8938,\n",
       " 'insight': 4839,\n",
       " 'archive': 1198,\n",
       " 'graphics': 4237,\n",
       " 'resources': 7641,\n",
       " 'list': 5476,\n",
       " 'modified': 5975,\n",
       " '1993': 211,\n",
       " '04': 7,\n",
       " '17': 138,\n",
       " 'computer': 2382,\n",
       " 'resource': 7640,\n",
       " 'listing': 5481,\n",
       " 'weekly': 9681,\n",
       " 'posting': 6949,\n",
       " 'april': 1183,\n",
       " 'faqs': 3729,\n",
       " 'including': 4734,\n",
       " 'available': 1383,\n",
       " 'site': 8263,\n",
       " 'pit': 6806,\n",
       " 'manager': 5658,\n",
       " 'mit': 5937,\n",
       " 'alias': 1021,\n",
       " 'rtfm': 7843,\n",
       " '172': 141,\n",
       " '27': 336,\n",
       " 'directory': 3081,\n",
       " 'pub': 7211,\n",
       " 'usenet': 9371,\n",
       " 'news': 6204,\n",
       " 'answers': 1122,\n",
       " 'archived': 1199,\n",
       " 'article': 1247,\n",
       " 'server': 8112,\n",
       " 'machine': 5609,\n",
       " 'send': 8077,\n",
       " 'message': 5834,\n",
       " 'containing': 2495,\n",
       " 'keyword': 5184,\n",
       " 'help': 4433,\n",
       " 'quotes': 7323,\n",
       " 'body': 1668,\n",
       " 'places': 6826,\n",
       " 'item': 5012,\n",
       " 'information': 4800,\n",
       " 'items': 5013,\n",
       " 'changed': 2050,\n",
       " 'arranged': 1234,\n",
       " 'subjects': 8641,\n",
       " 'order': 6465,\n",
       " 'limit': 5454,\n",
       " 'plan': 6830,\n",
       " 'changing': 2052,\n",
       " 'headers': 4402,\n",
       " 'soon': 8370,\n",
       " 'careful': 1942,\n",
       " 'keys': 5183,\n",
       " 'remain': 7546,\n",
       " 'subject': 8638,\n",
       " '3d': 446,\n",
       " 'editors': 3332,\n",
       " 'public': 7212,\n",
       " 'domain': 3187,\n",
       " 'free': 3986,\n",
       " 'shareware': 8160,\n",
       " 'systems': 8786,\n",
       " 'plotting': 6863,\n",
       " 'packages': 6554,\n",
       " 'molecular': 5983,\n",
       " 'visualization': 9535,\n",
       " 'stuff': 8630,\n",
       " 'thinking': 8966,\n",
       " 'making': 5650,\n",
       " 'bi': 1578,\n",
       " 'lines': 5465,\n",
       " 'character': 2058,\n",
       " 'added': 896,\n",
       " 'removed': 7564,\n",
       " 'diff': 3047,\n",
       " 'locate': 5513,\n",
       " 'changes': 2051,\n",
       " 'text': 8929,\n",
       " 'copyright': 2580,\n",
       " '1992': 210,\n",
       " 'copy': 2578,\n",
       " 'freely': 3989,\n",
       " 'provided': 7192,\n",
       " 'notice': 6268,\n",
       " 'intact': 4868,\n",
       " 'compiled': 2347,\n",
       " 'nick': 6222,\n",
       " 'nfotis': 6213,\n",
       " 'gr': 4218,\n",
       " 'contact': 2490,\n",
       " 'updates': 9344,\n",
       " 'corrections': 2602,\n",
       " 'disclaimer': 3100,\n",
       " 'guarantee': 4283,\n",
       " 'accuracy': 850,\n",
       " 'document': 3170,\n",
       " 'risk': 7752,\n",
       " 'mainly': 5635,\n",
       " 'guide': 4294,\n",
       " 'software': 8346,\n",
       " 'suggest': 8674,\n",
       " 'reading': 7417,\n",
       " 'comp': 2323,\n",
       " 'image': 4669,\n",
       " 'analysis': 1080,\n",
       " 'entitled': 3468,\n",
       " 'date': 2803,\n",
       " 'frequently': 3996,\n",
       " 'asked': 1259,\n",
       " 'john': 5074,\n",
       " 'jpl': 5102,\n",
       " 'nasa': 6122,\n",
       " 'poster': 6947,\n",
       " 'official': 6384,\n",
       " 'included': 4732,\n",
       " 'comments': 2305,\n",
       " 'contents': 2501,\n",
       " 'archie': 1196,\n",
       " 'notes': 6267,\n",
       " 'ftp': 4010,\n",
       " 'eric': 3496,\n",
       " 'servers': 8113,\n",
       " 'oriented': 6480,\n",
       " 'ray': 7394,\n",
       " 'tracing': 9090,\n",
       " 'related': 7520,\n",
       " 'mailing': 5631,\n",
       " 'lists': 5482,\n",
       " 'commercial': 2307,\n",
       " 'scene': 7958,\n",
       " 'description': 2963,\n",
       " 'languages': 5298,\n",
       " 'formats': 3941,\n",
       " 'processing': 7097,\n",
       " 'display': 3126,\n",
       " '11': 53,\n",
       " 'generators': 4109,\n",
       " 'geographical': 4121,\n",
       " 'data': 2800,\n",
       " 'maps': 5684,\n",
       " 'files': 3812,\n",
       " '12': 67,\n",
       " 'scanners': 7953,\n",
       " 'digitized': 3063,\n",
       " '13': 81,\n",
       " 'background': 1435,\n",
       " 'imagery': 4670,\n",
       " '14': 98,\n",
       " 'introduction': 4935,\n",
       " 'rendering': 7569,\n",
       " 'algorithms': 1019,\n",
       " 'buffer': 1793,\n",
       " 'depth': 2952,\n",
       " '15': 111,\n",
       " 'space': 8391,\n",
       " 'shuttle': 8212,\n",
       " '16': 125,\n",
       " 'scientific': 7976,\n",
       " '19': 160,\n",
       " 'future': 4039,\n",
       " 'additions': 901,\n",
       " 'info': 4796,\n",
       " 'crossposted': 2689,\n",
       " 'advised': 942,\n",
       " 'search': 8016,\n",
       " 'groups': 4272,\n",
       " 'usually': 9381,\n",
       " 'sysadmin': 8784,\n",
       " 'articles': 1248,\n",
       " 'posted': 6946,\n",
       " 'times': 9012,\n",
       " 'month': 5999,\n",
       " 'kept': 5175,\n",
       " 'field': 3795,\n",
       " 'inria': 4832,\n",
       " 'mirrors': 5919,\n",
       " 'port': 6919,\n",
       " '8000': 686,\n",
       " 'digest': 3060,\n",
       " 'type': 9227,\n",
       " 'purposes': 7240,\n",
       " 'place': 6823,\n",
       " 'monitors': 5995,\n",
       " 'robert': 7775,\n",
       " 'com': 2280,\n",
       " 'general': 4099,\n",
       " 'switch': 8759,\n",
       " 'swiss': 8758,\n",
       " 'academic': 820,\n",
       " 'network': 6192,\n",
       " 'switzerland': 8763,\n",
       " 'interactive': 4893,\n",
       " 'telnet': 8886,\n",
       " 'nic': 6219,\n",
       " 'ch': 2031,\n",
       " '130': 82,\n",
       " '59': 556,\n",
       " 'login': 5524,\n",
       " 'postings': 6950,\n",
       " 'typing': 9232,\n",
       " 'word': 9781,\n",
       " 'decide': 2845,\n",
       " 'following': 3908,\n",
       " 'methods': 5846,\n",
       " 'user': 9372,\n",
       " 'anonymous': 1117,\n",
       " 'internet': 4911,\n",
       " 'style': 8633,\n",
       " 'address': 903,\n",
       " 'password': 6645,\n",
       " 'cd': 1991,\n",
       " 'rfc': 7710,\n",
       " '400': 476,\n",
       " 'ou': 6499,\n",
       " 'receive': 7443,\n",
       " 'instructions': 4860,\n",
       " 'required': 7615,\n",
       " 'header': 4401,\n",
       " 'service': 8115,\n",
       " 'requested': 7611,\n",
       " 'appreciated': 1170,\n",
       " 'newsgroups': 6206,\n",
       " 'au': 1352,\n",
       " '139': 92,\n",
       " 'nz': 6310,\n",
       " 'funet': 4032,\n",
       " 'fi': 3792,\n",
       " '128': 77,\n",
       " '214': 277,\n",
       " 'finland': 3834,\n",
       " 'th': 8936,\n",
       " '83': 694,\n",
       " '111': 57,\n",
       " 'cs': 2708,\n",
       " 'ac': 819,\n",
       " 'il': 4660,\n",
       " '132': 85,\n",
       " '65': 598,\n",
       " 'jp': 5098,\n",
       " '54': 547,\n",
       " 'japan': 5035,\n",
       " 'kr': 5249,\n",
       " '163': 129,\n",
       " 'korea': 5244,\n",
       " 'tw': 9217,\n",
       " '140': 99,\n",
       " '115': 61,\n",
       " '24': 301,\n",
       " 'doc': 3163,\n",
       " 'ic': 4617,\n",
       " 'uk': 9267,\n",
       " '146': 106,\n",
       " '169': 135,\n",
       " 'ireland': 4976,\n",
       " 'net': 6186,\n",
       " '167': 133,\n",
       " '254': 320,\n",
       " '179': 148,\n",
       " 'usa': 9362,\n",
       " 'md': 5769,\n",
       " 'ne': 6148,\n",
       " '147': 107,\n",
       " '225': 288,\n",
       " 'ny': 6305,\n",
       " 'rutgers': 7866,\n",
       " 'nj': 6233,\n",
       " '195': 173,\n",
       " 'zealand': 9975,\n",
       " 'connect': 2439,\n",
       " 'username': 9373,\n",
       " 'clients': 2199,\n",
       " 'burden': 1816,\n",
       " 'session': 8118,\n",
       " 'x11': 9843,\n",
       " 'r5': 7337,\n",
       " 'contrib': 2524,\n",
       " 'tape': 8825,\n",
       " 'sources': 8385,\n",
       " 'misc': 5921,\n",
       " 'vol': 9550,\n",
       " 'account': 847,\n",
       " 'sites': 8264,\n",
       " 'note': 6265,\n",
       " 'janet': 5032,\n",
       " 'users': 9374,\n",
       " 'united': 9315,\n",
       " 'kingdom': 5210,\n",
       " 'accessible': 835,\n",
       " 'host': 4547,\n",
       " 'specify': 8421,\n",
       " 'messages': 5835,\n",
       " 'number': 6295,\n",
       " 'automated': 1371,\n",
       " 'things': 8964,\n",
       " 'response': 7652,\n",
       " 'email': 3392,\n",
       " 'query': 7303,\n",
       " 'unix': 9323,\n",
       " 'hours': 4557,\n",
       " 'days': 2815,\n",
       " 'later': 5316,\n",
       " '26': 329,\n",
       " '03': 5,\n",
       " '93': 726,\n",
       " 'eye': 3673,\n",
       " 'rayshade': 7397,\n",
       " 'tracer': 9089,\n",
       " 'workstations': 9793,\n",
       " 'mac': 5607,\n",
       " 'amiga': 1068,\n",
       " 'pov': 6961,\n",
       " 'son': 8366,\n",
       " 'trace': 9087,\n",
       " 'written': 9825,\n",
       " 'drew': 3232,\n",
       " 'wells': 9688,\n",
       " 'compuserve': 2377,\n",
       " 'dave': 2808,\n",
       " 'buck': 1787,\n",
       " 'ca': 1864,\n",
       " 'art': 1245,\n",
       " 'range': 7371,\n",
       " 'surface': 8721,\n",
       " 'types': 9229,\n",
       " 'package': 6553,\n",
       " 'pcs': 6678,\n",
       " 'ii': 4654,\n",
       " 'vms': 9543,\n",
       " 'previewer': 7042,\n",
       " 'rtrace': 7844,\n",
       " 'patches': 6651,\n",
       " 'ms': 6058,\n",
       " 'dos': 3199,\n",
       " 'version': 9477,\n",
       " 'exists': 3608,\n",
       " 'binary': 1599,\n",
       " '286': 348,\n",
       " '287': 349,\n",
       " 'author': 1363,\n",
       " 'stephen': 8551,\n",
       " 'vax': 9451,\n",
       " 'boeing': 1669,\n",
       " '386': 437,\n",
       " 'source': 8384,\n",
       " 'registered': 7504,\n",
       " '50': 532,\n",
       " 'direct': 3073,\n",
       " 'steve': 8555,\n",
       " 'dimensional': 3064,\n",
       " 'separate': 8095,\n",
       " 'wireframe': 9743,\n",
       " 'viewer': 9501,\n",
       " 'mtv': 6066,\n",
       " 'interesting': 4896,\n",
       " 'features': 3760,\n",
       " 'distributed': 3144,\n",
       " 'parallel': 6598,\n",
       " 'binaries': 1598,\n",
       " 'hp': 4565,\n",
       " 'series': 8107,\n",
       " 'needs': 6167,\n",
       " 'sun': 8693,\n",
       " 'newer': 6200,\n",
       " 'id': 4624,\n",
       " 'volume': 9554,\n",
       " 'cornell': 2589,\n",
       " 'renderer': 7568,\n",
       " 'warren': 9636,\n",
       " 'fortran': 3951,\n",
       " 'radiosity': 7355,\n",
       " 'lighting': 5442,\n",
       " 'radiance': 7350,\n",
       " 'effects': 3350,\n",
       " 'greg': 4256,\n",
       " 'ward': 9625,\n",
       " 'excellent': 3573,\n",
       " 'shading': 8142,\n",
       " 'models': 5965,\n",
       " 'physically': 6774,\n",
       " 'based': 1481,\n",
       " 'simulation': 8250,\n",
       " 'ported': 6922,\n",
       " 'india': 4755,\n",
       " 'indian': 4756,\n",
       " 'sgi': 8138,\n",
       " 'machines': 5610,\n",
       " 'includes': 4733,\n",
       " 'house': 4558,\n",
       " 'database': 2801,\n",
       " 'solution': 8357,\n",
       " 'run': 7854,\n",
       " 'stand': 8497,\n",
       " 'box': 1717,\n",
       " 'walk': 9611,\n",
       " 'requires': 7618,\n",
       " '4d': 514,\n",
       " 'libraries': 5421,\n",
       " 'scan': 7950,\n",
       " 'uses': 9375,\n",
       " 'shadow': 8143,\n",
       " 'algorithm': 1018,\n",
       " 'tcl': 8844,\n",
       " 'command': 2298,\n",
       " 'interface': 4899,\n",
       " 'program': 7116,\n",
       " 'set': 8120,\n",
       " 'commands': 2301,\n",
       " 'programmed': 7118,\n",
       " 'write': 9819,\n",
       " 'compile': 2346,\n",
       " 'code': 2239,\n",
       " 'surfaces': 8722,\n",
       " 'objects': 6331,\n",
       " 'scenes': 7959,\n",
       " 'ppm': 6972,\n",
       " 'format': 3939,\n",
       " 'utah': 9383,\n",
       " 'raster': 7384,\n",
       " 'toolkit': 9056,\n",
       " 'rle': 7762,\n",
       " 'photo': 6766,\n",
       " 'widget': 9709,\n",
       " 'tk': 9027,\n",
       " 'applications': 1162,\n",
       " 'learning': 5364,\n",
       " 'environment': 3476,\n",
       " 'device': 3016,\n",
       " 'portable': 6920,\n",
       " 'library': 5422,\n",
       " 'fast': 3736,\n",
       " 'polygon': 6905,\n",
       " 'intel': 4875,\n",
       " 'dr': 3216,\n",
       " 'journal': 5092,\n",
       " 'budget': 1790,\n",
       " 'texture': 8931,\n",
       " 'mapping': 5683,\n",
       " 'viewers': 9502,\n",
       " 'vision': 9529,\n",
       " 'output': 6508,\n",
       " 'solid': 8356,\n",
       " 'support': 8707,\n",
       " 'major': 5642,\n",
       " 'archives': 1200,\n",
       " 'wuarchive': 9833,\n",
       " 'zip': 9984,\n",
       " 'povray': 6963,\n",
       " 'beta': 1569,\n",
       " 'french': 3992,\n",
       " 'docs': 3165,\n",
       " '250': 316,\n",
       " 'save': 7936,\n",
       " 'disabled': 3086,\n",
       " 'extra': 3667,\n",
       " 'utilities': 9387,\n",
       " 'dxf': 3278,\n",
       " 'et': 3527,\n",
       " 'al': 1006,\n",
       " 'mathematical': 5740,\n",
       " 'structures': 8618,\n",
       " 'looking': 5534,\n",
       " 'geometry': 4122,\n",
       " 'minnesota': 5905,\n",
       " 'experimental': 3625,\n",
       " 'computation': 2378,\n",
       " 'macintosh': 5611,\n",
       " 'computers': 2383,\n",
       " 'sets': 8122,\n",
       " 'spd': 8401,\n",
       " 'databases': 2802,\n",
       " 'testing': 8925,\n",
       " 'nff': 6212,\n",
       " 'imagine': 4673,\n",
       " 'converters': 2551,\n",
       " 'nice': 6220,\n",
       " 'postscript': 6952,\n",
       " 'object': 6325,\n",
       " 'converts': 2554,\n",
       " 'outputs': 6509,\n",
       " 'views': 9505,\n",
       " 'tex': 8927,\n",
       " 'pk': 6819,\n",
       " 'font': 3912,\n",
       " 'converter': 2550,\n",
       " 'generator': 4108,\n",
       " 'glenn': 4168,\n",
       " 'lewis': 5403,\n",
       " 'hill': 4469,\n",
       " 'heads': 4405,\n",
       " 'brain': 1730,\n",
       " 'knee': 5228,\n",
       " 'electron': 3378,\n",
       " 'density': 2939,\n",
       " 'material': 5735,\n",
       " 'rt': 7842,\n",
       " 'collections': 2262,\n",
       " 'references': 7480,\n",
       " 'refer': 7478,\n",
       " 'global': 4169,\n",
       " 'rick': 7723,\n",
       " 'cross': 2687,\n",
       " 'abstracts': 812,\n",
       " 'collection': 2261,\n",
       " 'tom': 9044,\n",
       " 'wilson': 9720,\n",
       " 'paper': 6590,\n",
       " 'bank': 1464,\n",
       " 'project': 7129,\n",
       " 'various': 9445,\n",
       " 'technical': 8859,\n",
       " 'papers': 6592,\n",
       " 'form': 3937,\n",
       " 'online': 6418,\n",
       " 'bibliography': 1583,\n",
       " 'acm': 865,\n",
       " 'siggraph': 8220,\n",
       " '000': 1,\n",
       " 'unique': 9313,\n",
       " 'computational': 2379,\n",
       " 'community': 2322,\n",
       " 'educational': 3340,\n",
       " 'located': 5514,\n",
       " 'org': 6471,\n",
       " 'download': 3208,\n",
       " 'log': 5520,\n",
       " 'entries': 3471,\n",
       " 'access': 833,\n",
       " 'followed': 3906,\n",
       " 'topic': 9059,\n",
       " 'wish': 9753,\n",
       " 'index': 4754,\n",
       " 'suggestions': 8678,\n",
       " 'directed': 3074,\n",
       " 'admin': 915,\n",
       " 'manipulation': 5667,\n",
       " 'tools': 9058,\n",
       " 'pbmplus': 6674,\n",
       " 'conversion': 2547,\n",
       " 'writing': 9823,\n",
       " 'tiff': 9003,\n",
       " 'images': 4671,\n",
       " 'xv': 9898,\n",
       " 'xloadimage': 9875,\n",
       " 'displays': 3129,\n",
       " 'screen': 7991,\n",
       " 'khoros': 5190,\n",
       " 'huge': 4574,\n",
       " 'visual': 9534,\n",
       " 'programming': 7121,\n",
       " 'windows': 9726,\n",
       " 'somewhat': 8365,\n",
       " 'bit': 1613,\n",
       " 'plays': 6854,\n",
       " 'animation': 1102,\n",
       " 'resolution': 7635,\n",
       " 'gif': 4148,\n",
       " 'extensions': 3659,\n",
       " 'dl': 3158,\n",
       " 'pictures': 6788,\n",
       " 'ham': 4330,\n",
       " 'color': 2268,\n",
       " 'san': 7913,\n",
       " 'diego': 3043,\n",
       " 'paint': 6570,\n",
       " '24bit': 310,\n",
       " 'useful': 9369,\n",
       " 'books': 1683,\n",
       " 'patch': 6650,\n",
       " 'tar': 8828,\n",
       " 'intersection': 4921,\n",
       " 'routines': 7827,\n",
       " 'sean': 8015,\n",
       " 'graves': 4243,\n",
       " 'uniform': 9310,\n",
       " 'properties': 7154,\n",
       " '75': 650,\n",
       " 'har': 4360,\n",
       " 'el': 3369,\n",
       " 'rl': 7760,\n",
       " 'means': 5779,\n",
       " 'distributor': 3146,\n",
       " 'north': 6260,\n",
       " 'america': 1063,\n",
       " 'continent': 2505,\n",
       " 'wustl': 9834,\n",
       " '252': 318,\n",
       " '135': 88,\n",
       " 'msdos': 6059,\n",
       " ...}"
      ]
     },
     "execution_count": 77,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vocabulary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "wW6-QiaOLKzO"
   },
   "outputs": [],
   "source": [
    "docs = []\n",
    "for row in tf.toarray():\n",
    "    present_words = np.where(row != 0)[0].tolist()\n",
    "    present_words_with_count = []\n",
    "    for word_idx in present_words:\n",
    "        for count in range(row[word_idx]):\n",
    "            present_words_with_count.append(word_idx)\n",
    "    docs.append(present_words_with_count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "6GNG9Z_dLMzr"
   },
   "outputs": [],
   "source": [
    "inv_vocabulary = {v: k for k, v in vocabulary.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 157
    },
    "id": "QylY_rdnLPDx",
    "outputId": "67d6f95d-599c-4eb1-a2e1-63d025a742f0"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.google.colaboratory.intrinsic+json": {
       "type": "string"
      },
      "text/plain": [
       "\"Well i'm not sure about the story nad it did seem biased. What\\nI disagree with is your statement that the U.S. Media is out to\\nruin Israels reputation. That is rediculous. The U.S. media is\\nthe most pro-israeli media in the world. Having lived in Europe\\nI realize that incidences such as the one described in the\\nletter have occured. The U.S. media as a whole seem to try to\\nignore them. The U.S. is subsidizing Israels existance and the\\nEuropeans are not (at least not to the same degree). So I think\\nthat might be a reason they report more clearly on the\\natrocities.\\n\\tWhat is a shame is that in Austria, daily reports of\\nthe inhuman acts commited by Israeli soldiers and the blessing\\nreceived from the Government makes some of the Holocaust guilt\\ngo away. After all, look how the Jews are treating other races\\nwhen they got power. It is unfortunate.\\n\""
      ]
     },
     "execution_count": 80,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "920L7ELMLQ6H",
    "outputId": "c6047206-d046-4bce-ee9a-4f1471556f1f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acts atrocities austria away biased clearly daily degree described did disagree europe europeans existance got government guilt having holocaust ignore israeli israeli jews letter lived look makes media media media media occured power pro races realize reason received report reports reputation shame soldiers statement story sure think treating try unfortunate world "
     ]
    }
   ],
   "source": [
    "for i in docs[0]:\n",
    "  print(inv_vocabulary[i], end=\" \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ETjd-y3OLYrI"
   },
   "outputs": [],
   "source": [
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "JY_3sGHrLZ-F"
   },
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(plt, cm,\n",
    "                          classes,\n",
    "                          clusters,\n",
    "                          title = ''):\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=plt.cm.Blues)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    xtick_marks = np.arange(len(clusters))\n",
    "    ytick_marks = np.arange(len(classes))\n",
    "    plt.xticks(xtick_marks, clusters, rotation=45)\n",
    "    plt.yticks(ytick_marks, classes)\n",
    "    plt.xlabel('Predicted Label')\n",
    "    plt.ylabel('True Label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "kWLqIK1ILbZf"
   },
   "outputs": [],
   "source": [
    "def purity_score(y_true, y_pred):\n",
    "    # compute contingency matrix (also called confusion matrix)\n",
    "    contingency_matrix = metrics.cluster.contingency_matrix(y_true, y_pred)\n",
    "    # return purity\n",
    "    return np.sum(np.amax(contingency_matrix, axis=0)) / np.sum(contingency_matrix) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "juMutK50KyyC"
   },
   "source": [
    "## LDA using custom functions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "FTd10-xdK5AZ"
   },
   "source": [
    "### Gibbs sampling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "OMpBl9wrK4LZ"
   },
   "outputs": [],
   "source": [
    "M = len(docs)  # number of documents in the collection\n",
    "N = len(vocabulary)  # number of words in the collection\n",
    "K = 20  # number of topics\n",
    "\n",
    "z = np.zeros(shape=(N))  # topic assignment for each word\n",
    "n_d_k = np.zeros(shape=(M, K))  # Number of times document d uses topic k\n",
    "n_k_w = np.zeros(shape=(K, N))  # Number of times topic k uses word type wd,n\n",
    "\n",
    "beta = 1 / K         # the parameter of the Dirichlet prior on the per-document topic distributions\n",
    "alpha = 1 / K        # the parameter of the Dirichlet prior on the per-topic word distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 67,
     "referenced_widgets": [
      "d39aa5ddaafa4c7a8efca9a7123357bc",
      "227be25701ec4e92b20f475a858b854f",
      "1fbe6bef23184eecb7efb24036196bdb",
      "3277a95c4b1d4a278fae78f99f5b42eb",
      "618c789eccd1447db7eaab21aae7da9b",
      "6e9f1e99023b4224b83e98dea1fe84f3",
      "4b9af6a945ee47f982b362702fd60327",
      "00d59c369ac443d89786ef4b8040ac7a"
     ]
    },
    "id": "Fwy2DFybK7zj",
    "outputId": "a38ae96c-fd05-4e1c-9cad-1b95b7791b42"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d39aa5ddaafa4c7a8efca9a7123357bc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "z = []\n",
    "for ind in tqdm(range(1)):\n",
    "  for docIdx, doc in enumerate(docs):\n",
    "    # print(doc)\n",
    "    z_d = []\n",
    "    for wordIdx in doc:\n",
    "      topicIdx = len(doc) % K\n",
    "      z_d.append(topicIdx)\n",
    "      n_d_k[docIdx][topicIdx] += 1\n",
    "      n_k_w[topicIdx][wordIdx] += 1\n",
    "    z.append(z_d)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 67,
     "referenced_widgets": [
      "69a349e5087142ad8132e303e499b9d6",
      "e99b07f70b8b4740abfc29b53cb99617",
      "6885b492588342afb1403a7c0bce99af",
      "8da1c14f46ab4e0cb1d166636d65ca7d",
      "7dbbf0e6e62d4ae3875b20dfee860d14",
      "1a00528558194b6b893ba88548263173",
      "87aa2524714942639f2317d2f11b4d9f",
      "0e34c7826bc34425ac785df9d50b0fab"
     ]
    },
    "id": "JI7cmqssLlNx",
    "outputId": "a75892be-d521-4d11-9d8a-d1b6566d9cfd"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "69a349e5087142ad8132e303e499b9d6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=10.0), HTML(value='')))"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "for iter in tqdm(range(10)):\n",
    "  for docIdx, doc in enumerate(docs):\n",
    "    # print(\"Initial n_d_k\", n_d_k[docIdx])\n",
    "    for w, wordIdx in enumerate(doc):\n",
    "        # Fetch topic of the word\n",
    "        topicIdx = z[docIdx][w]\n",
    "\n",
    "        # Decrement counters corresponding to this word\n",
    "        n_d_k[docIdx][topicIdx] -= 1\n",
    "        n_k_w[topicIdx][wordIdx] -= 1\n",
    "\n",
    "\n",
    "        p_d_k = (n_d_k[docIdx] + alpha) / (np.sum(n_d_k) + K * alpha)\n",
    "        p_k_w = (n_k_w[:, wordIdx] + beta) / (np.sum(n_k_w) + N*beta)\n",
    "        p_z = p_d_k * p_k_w\n",
    "        p_z = p_z / np.sum(p_z)\n",
    "        # print(p_z)\n",
    "\n",
    "        newTopicIdx = np.random.multinomial(1, p_z).argmax()\n",
    "        z[docIdx][w] = newTopicIdx\n",
    "\n",
    "        # Increment counters corresponding to this word\n",
    "        n_d_k[docIdx][newTopicIdx] += 1\n",
    "        n_k_w[newTopicIdx][wordIdx] += 1\n",
    "      # print(\"Final n_d_k\", n_d_k[docIdx])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "3qZeMwtQLqMT"
   },
   "outputs": [],
   "source": [
    "def plot_topic_dist(word):\n",
    "  ind = np.where(np.asarray(list(vocabulary.keys()))==word)[0][0]\n",
    "  vInd = list(vocabulary.values())[ind]\n",
    "  vWord = list(vocabulary.keys())[ind]\n",
    "  plt.plot(range(20), n_k_w.T[ind]/ sum(n_k_w.T[ind]))\n",
    "  plt.title(\"Topic distribution for word: '{}'\".format(vWord))\n",
    "  plt.xticks(range(20))\n",
    "  plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "id": "dZDJsTxSLsSD",
    "outputId": "f1214f93-7f47-47b4-8e2d-fd953638dbee"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_topic_dist(\"religion\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "id": "XFvO108CpPzS",
    "outputId": "15aa1d77-691e-435f-8b76-5e0c94f8e3f5"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_topic_dist(\"christ\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "jsJ57SuFLw7X",
    "outputId": "c60f6bb5-f73c-4b64-a32c-478f620b5f87"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Topic #0: w7 uw t7 chz cx ck w1 hz a7 lk\n",
      "Topic #1: people don know said just like think did say right\n",
      "Topic #2: chastity geb dsl n3jxp shameful cadre intellect skepticism pitt gordon\n",
      "Topic #3: astronaut larson directive bobbe unified 70s physicists catcher queens pitched\n",
      "Topic #4: db mov bh si byte bl di al cs cx\n",
      "Topic #5: cx c_ s6 ck hz lk chz sc_ w7 17\n",
      "Topic #6: like use edu just don know new does 10 time\n",
      "Topic #7: dod polio dole dee ax bolshevik 1d9 3t 9v chastity\n",
      "Topic #8: zz edition edu edt edmonton edm editors editorial editor editing\n",
      "Topic #9: ax max g9v b8f a86 pl 1d9 3t 145 1t\n",
      "Topic #10: armenian armenians jews azerbaijan turkish don ottoman did dod armenia\n",
      "Topic #11: ax b8f 145 max a86 0t pl 34u 1t ah\n",
      "Topic #12: people just don like think know time use does god\n",
      "Topic #13: spine inference cmap 1948 1964 docs editorial edu edt edmonton\n",
      "Topic #14: cubs suck hiv vaccines infectious hampshire clinical fundamentalists billboard protest\n",
      "Topic #15: ax max 75u 2di 145 0d 34u sl b8f bhj\n",
      "Topic #16: grounding conductor gfci candida cylinders bios jumper ground utica conductors\n",
      "Topic #17: eff sleeve 250 stephanopoulos promo cooper cud enforcement abs deaf\n",
      "Topic #18: een 2nd burns fits edges edit edited editing edition editor\n",
      "Topic #19: cfg latin ______ zz edition edu edt edmonton edm editors\n"
     ]
    }
   ],
   "source": [
    "n_top_words = 10\n",
    "for topic_idx, topic in enumerate(n_k_w):\n",
    "    message = \"Topic #%d: \" % topic_idx\n",
    "    message += \" \".join([inv_vocabulary[i] for i in topic.argsort()[:-n_top_words - 1:-1]])\n",
    "    print(message)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "5yl2X90Kcw9E"
   },
   "outputs": [],
   "source": [
    "predLabels = []\n",
    "for i in n_d_k:\n",
    "  predLabels.append(np.argmax(i))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Me3mbEK9dSNb"
   },
   "outputs": [],
   "source": [
    "with open('/content/drive/MyDrive/Colab Notebooks/predLabels_Gibbs_custom.pkl', 'wb') as f:\n",
    "  pickle.dump(predLabels, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "5gZ2wBUYgD8Y"
   },
   "outputs": [],
   "source": [
    "file = open('/content/drive/MyDrive/Colab Notebooks/predLabels_Gibbs_custom.pkl','rb')\n",
    "predLabels = pickle.load(file)\n",
    "file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 331
    },
    "id": "ypIOc0tcY0NI",
    "outputId": "661c4386-724f-4dcb-9350-1d690719411a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "clusterNames = [\"Cluster%d\" % i for i in range(20)]\n",
    "targetNames = [\"Target%d\" % i for i in range(20)]\n",
    "cm = metrics.confusion_matrix(target, predLabels)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(plt, cm,\n",
    "                       classes=targetNames,\n",
    "                       clusters=clusterNames,\n",
    "                       title='LDA with custom Gibbs sampling')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "EkwmVtQUcnlb",
    "outputId": "ff482900-c135-4bba-8ec8-3ebce52bbd6b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LDA with custom Gibbs sampling\n",
      "Homogeneity 0.04063416533673405\n",
      "Completenesee 0.11067726730802309\n",
      "Normalized Mutual Information 0.05944399970583384\n",
      "Purity 0.09395439278769666\n"
     ]
    }
   ],
   "source": [
    "print(\"LDA with custom Gibbs sampling\")\n",
    "print(\"Homogeneity\", metrics.homogeneity_score(target, predLabels))\n",
    "print(\"Completenesee\", metrics.completeness_score(target, predLabels))\n",
    "print(\"Normalized Mutual Information\", metrics.normalized_mutual_info_score(target, predLabels))\n",
    "print(\"Purity\", purity_score(target, predLabels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "g5Swjdy8N6Hq"
   },
   "source": [
    "### Variational inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "CORPdL5NN8Dt"
   },
   "outputs": [],
   "source": [
    "from scipy.special import digamma, polygamma"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "rXmNYnJXOCwn",
    "outputId": "8b575d08-78f5-4ae1-87c1-1f9cabae6504"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import toolz as tz\n",
    "import toolz.curried as c\n",
    "import string\n",
    "from nltk.corpus import stopwords\n",
    "import nltk\n",
    "nltk.download('stopwords')\n",
    "import re\n",
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "KE4Bvg_2OGf0"
   },
   "outputs": [],
   "source": [
    "part = 1000\n",
    "d_sub = docs[:part]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "iHYp4HDJOMnf"
   },
   "outputs": [],
   "source": [
    "tf = {id: tz.frequencies(doc) for id, doc in enumerate(d_sub)}\n",
    "df = pd.DataFrame(tf).fillna(0)\n",
    "words = df.index\n",
    "\n",
    "ds = df.values.T\n",
    "ds = ds.astype(int)\n",
    "\n",
    "def DataTrans(x):\n",
    "    \"\"\"Turn the data into the desired structure\"\"\"\n",
    "    \n",
    "    N_d = np.count_nonzero(x)\n",
    "    V = len(x)\n",
    "    \n",
    "    row = 0\n",
    "    \n",
    "    doc = np.zeros((N_d, V))\n",
    "    for i in range(V):\n",
    "        if x[i] == 0:\n",
    "            pass\n",
    "        else:\n",
    "            doc[row, i] = x[i]\n",
    "            row += 1\n",
    "    \n",
    "    return doc\n",
    "\n",
    "docsTransformed = list(map(DataTrans, ds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "gUy7Z7UNOW6h",
    "outputId": "05967ba5-0a8b-4faa-f22f-35d0c227b300"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "M 1000\n",
      "V 7930\n",
      "N [47, 38, 35, 62, 31, 16, 114, 40, 27, 56, 1514, 19, 18, 37, 8, 10, 47, 18, 10, 98, 37, 48, 35, 39, 95, 78, 17, 79, 48, 16, 180, 28, 63, 52, 240, 10, 28, 31, 40, 12, 109, 32, 75, 17, 1, 124, 7, 24, 69, 57, 41, 0, 31, 9, 36, 194, 10, 64, 2, 56, 86, 34, 10, 23, 30, 10, 9, 23, 38, 53, 48, 3, 22, 20, 22, 122, 7, 42, 94, 171, 79, 20, 11, 20, 17, 43, 68, 69, 28, 0, 26, 98, 15, 57, 56, 45, 11, 28, 21, 78, 58, 29, 26, 50, 28, 13, 15, 17, 68, 27, 3, 11, 92, 9, 15, 84, 54, 55, 107, 15, 19, 30, 68, 12, 22, 30, 73, 80, 59, 44, 40, 42, 75, 53, 37, 103, 42, 32, 53, 16, 17, 10, 58, 15, 41, 10, 93, 9, 96, 29, 0, 49, 33, 18, 31, 3, 29, 44, 45, 10, 60, 55, 41, 33, 15, 102, 27, 68, 47, 7, 31, 49, 36, 37, 43, 17, 8, 20, 50, 69, 57, 10, 30, 28, 21, 47, 12, 20, 72, 27, 15, 19, 133, 13, 39, 52, 35, 12, 31, 9, 836, 145, 27, 20, 47, 38, 10, 120, 9, 21, 34, 7, 33, 42, 70, 30, 16, 64, 59, 3, 83, 80, 56, 34, 37, 45, 26, 25, 8, 8, 57, 52, 5, 44, 12, 28, 16, 38, 134, 77, 40, 48, 42, 3, 55, 11, 61, 26, 19, 20, 17, 27, 53, 40, 47, 0, 17, 398, 2, 79, 80, 37, 18, 55, 1, 55, 8, 23, 222, 22, 17, 40, 31, 30, 60, 26, 44, 24, 38, 47, 37, 75, 9, 24, 228, 29, 18, 45, 47, 58, 441, 29, 12, 26, 15, 66, 63, 159, 18, 29, 42, 3, 3, 35, 51, 12, 27, 18, 149, 27, 21, 0, 72, 211, 51, 34, 20, 18, 53, 20, 21, 83, 2, 3, 417, 17, 53, 12, 94, 27, 19, 13, 237, 68, 19, 87, 13, 19, 0, 85, 8, 47, 33, 8, 55, 15, 27, 66, 11, 22, 444, 3, 16, 22, 183, 147, 53, 62, 2, 0, 64, 0, 46, 35, 206, 52, 40, 71, 9, 129, 70, 82, 26, 82, 37, 14, 15, 29, 199, 16, 47, 46, 19, 25, 78, 11, 11, 234, 26, 16, 18, 50, 27, 26, 65, 50, 22, 50, 23, 35, 78, 17, 0, 26, 164, 4, 49, 42, 0, 76, 203, 64, 101, 18, 59, 94, 8, 106, 31, 41, 35, 41, 27, 55, 35, 16, 22, 6, 9, 18, 542, 42, 0, 21, 99, 13, 62, 44, 62, 0, 0, 59, 27, 20, 32, 18, 41, 31, 27, 38, 55, 58, 35, 162, 8, 76, 27, 31, 31, 162, 27, 28, 58, 14, 72, 21, 8, 9, 24, 34, 16, 241, 134, 58, 55, 62, 112, 87, 53, 12, 46, 21, 0, 0, 6, 15, 27, 43, 5, 25, 22, 15, 14, 58, 4, 19, 167, 44, 106, 71, 0, 3, 38, 30, 20, 14, 21, 49, 37, 24, 154, 37, 24, 12, 24, 273, 171, 48, 110, 24, 8, 7, 20, 41, 53, 99, 15, 34, 4, 27, 28, 31, 45, 16, 8, 208, 100, 19, 49, 42, 25, 69, 18, 17, 77, 8, 17, 18, 68, 182, 84, 19, 0, 11, 58, 90, 12, 13, 29, 2, 13, 108, 14, 0, 4, 171, 55, 37, 73, 26, 17, 18, 19, 38, 9, 23, 12, 34, 20, 41, 28, 42, 33, 68, 24, 21, 63, 83, 29, 430, 62, 75, 31, 87, 31, 6, 15, 95, 0, 1, 47, 9, 179, 11, 2, 40, 10, 45, 26, 11, 42, 11, 29, 9, 23, 75, 28, 43, 405, 56, 139, 36, 21, 41, 30, 31, 11, 32, 21, 18, 39, 19, 45, 34, 36, 48, 46, 11, 18, 116, 65, 8, 30, 72, 90, 28, 21, 103, 14, 86, 32, 21, 60, 15, 89, 41, 16, 70, 26, 24, 178, 72, 16, 19, 0, 67, 62, 45, 60, 91, 11, 8, 29, 16, 47, 31, 16, 18, 24, 229, 17, 4, 18, 101, 28, 95, 69, 55, 25, 14, 20, 11, 32, 166, 28, 34, 8, 35, 26, 17, 38, 28, 44, 178, 28, 48, 53, 24, 26, 96, 15, 257, 15, 28, 93, 27, 42, 9, 49, 163, 9, 21, 33, 3, 69, 17, 16, 120, 5, 67, 78, 80, 0, 77, 154, 24, 15, 16, 58, 36, 38, 39, 88, 63, 29, 62, 37, 15, 53, 43, 201, 289, 69, 18, 7, 14, 17, 71, 41, 57, 50, 54, 62, 133, 0, 41, 74, 37, 71, 46, 2, 29, 93, 43, 64, 17, 14, 61, 27, 4, 23, 48, 89, 22, 254, 10, 91, 41, 12, 16, 28, 8, 38, 15, 3, 13, 41, 19, 19, 155, 23, 27, 16, 6, 25, 35, 665, 32, 149, 0, 14, 15, 54, 28, 40, 7, 24, 83, 22, 22, 61, 13, 42, 87, 74, 8, 27, 78, 103, 11, 43, 65, 76, 32, 380, 114, 15, 35, 14, 40, 21, 21, 63, 67, 36, 32, 98, 61, 58, 26, 42, 30, 17, 41, 29, 52, 64, 79, 25, 124, 10, 37, 67, 18, 56, 39, 0, 317, 75, 17, 131, 22, 3, 17, 23, 91, 18, 4, 88, 144, 29, 39, 12, 10, 31, 13, 24, 30, 23, 48, 47, 34, 11, 36, 58, 75, 50, 13, 40, 13, 8, 29, 21, 26, 12, 0, 44, 31, 13, 18, 32, 6, 0, 226, 0, 19, 31, 23, 85, 23, 544, 34, 66, 37, 32, 14, 42, 27, 46, 44, 33, 45, 89, 101, 21, 15, 27, 3, 27, 16, 28, 56, 8, 25, 66, 25, 32, 27, 3, 50, 4, 53, 15, 35, 4, 10, 53, 95, 48, 19, 33, 14, 15, 181, 50, 4, 14, 7, 12, 31, 16, 18, 31, 62, 58, 0, 34, 54, 11, 210, 28, 7, 13, 15, 48, 2, 11, 7, 7, 18, 27, 11, 47, 56, 29, 45, 51, 24, 59, 51]\n"
     ]
    }
   ],
   "source": [
    "# docs = docsTransformed\n",
    "M = len(docsTransformed)\n",
    "V = docsTransformed[1].shape[1]\n",
    "N = [doc.shape[0] for doc in docsTransformed]\n",
    "K = 20\n",
    "\n",
    "initial_alpha_shape = 100\n",
    "initial_alpha_scale = 0.01\n",
    "\n",
    "print(\"M\", M)\n",
    "print(\"V\", V)\n",
    "print(\"N\", N)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "PtrlCUZBOYkh",
    "outputId": "18797500-1245-487b-e479-d007c4865520"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "beta0 (20, 7930)\n",
      "alpha0 (20,)\n",
      "phi0 (47, 20)\n",
      "gamma (1000, 20)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:5: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "  \"\"\"\n"
     ]
    }
   ],
   "source": [
    "# initialization\n",
    "np.random.seed(100)\n",
    "beta0 = np.random.dirichlet(np.ones(V), K)\n",
    "alpha0 = np.random.gamma(shape=initial_alpha_shape, scale=initial_alpha_scale, size=K)\n",
    "phi0 = np.array([np.ones((N[d],K))/K for d in range(M)])\n",
    "gamma0 = np.array([alpha0+N[d]/K for d in range(M)])\n",
    "\n",
    "# BETA0 = np.zeros(shape=(k, V))\n",
    "# alpha0 = np.asarray([50/k]*k)\n",
    "# PHI = np.asarray([1/k]*M)\n",
    "# GAMMA = np.array([alpha0 + N[d]/k for d in range(M)])\n",
    "\n",
    "print(\"beta0\", beta0.shape)\n",
    "print(\"alpha0\", alpha0.shape)\n",
    "print(\"phi0\", phi0[0].shape)\n",
    "print(\"gamma\", gamma0.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "RtIJiGYZOaUl"
   },
   "outputs": [],
   "source": [
    "def e_step(doc, alpha, beta, phi0, gamma0):\n",
    "  tol = 1e-3**2\n",
    "  #for docIdx, doc in enumerate(docs):\n",
    "    # print(\"Doc\", docIdx, doc.shape)\n",
    "    # print(\"phi[docIdx]\", phi[docIdx].shape)\n",
    "  \n",
    "  phi = phi0\n",
    "  gamma = gamma0\n",
    "\n",
    "  phi_delta = 1\n",
    "  gamma_delta = 1\n",
    "\n",
    "  phi = phi.astype(np.float128)\n",
    "  phi0 = phi0.astype(np.float128)\n",
    "\n",
    "  for iterations in range(100):\n",
    "      # compute new values\n",
    "      gamma = gamma / min(gamma)\n",
    "      phi = (np.matmul(doc, beta.T)) * np.exp(digamma(gamma)-digamma(sum(gamma)))\n",
    "      # print(\"phi[docIdx]\", phi[docIdx].shape)\n",
    "      phi = phi / (phi.sum(axis=1).reshape(-1,1))   \n",
    "      \n",
    "      gamma = alpha + phi.sum(axis=0)\n",
    "\n",
    "      # compute delta\n",
    "      phi_delta = np.mean((phi - phi0) ** 2)\n",
    "      gamma_delta = np.mean((gamma - gamma0) ** 2)\n",
    "\n",
    "      phi0 = phi\n",
    "      gamma0 = gamma\n",
    "\n",
    "      if phi_delta <= tol and gamma_delta <= tol:\n",
    "        break\n",
    "\n",
    "  return phi, gamma"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "2YNBVNvcObxl"
   },
   "outputs": [],
   "source": [
    "def m_step(docs, alpha0, beta0, phi0, gamma0):\n",
    "  tol = 1e-3**2\n",
    "\n",
    "  alpha = alpha0\n",
    "  beta = beta0\n",
    "\n",
    "  alpha_delta = 1\n",
    "  beta_delta = 1\n",
    "\n",
    "  phi = phi0\n",
    "  gamma = gamma0\n",
    "  \n",
    "  for iter in tqdm(range(100)):\n",
    "    # update beta\n",
    "    beta = np.zeros(shape=(K, V))\n",
    "    for docIdx, doc in enumerate(docs):\n",
    "      phi[docIdx], gamma[docIdx,] = e_step(doc, alpha0, beta0, phi[docIdx], gamma[docIdx,])\n",
    "      beta += np.matmul(phi0[docIdx].T, doc)\n",
    "    beta = beta / beta.sum(axis=1).reshape(-1,1)\n",
    "\n",
    "    # update alpha\n",
    "    z = M * polygamma(1, sum(alpha0))\n",
    "    h = -M * polygamma(1, alpha0)\n",
    "    g = M * (digamma(sum(alpha0)) - digamma(alpha0)) + (digamma(gamma) - digamma(gamma.sum(axis=1).reshape(-1,1))).sum(axis=0)\n",
    "    c = (sum(g/h)) / (1/z + sum(1/h))\n",
    "    alpha = alpha0 - (g-c)/h\n",
    "\n",
    "    alpha_delta = np.mean((alpha - alpha0)**2)\n",
    "    beta_delta = np.mean((beta - beta0)**2)\n",
    "\n",
    "    alpha0 = alpha\n",
    "    beta0 = beta\n",
    "\n",
    "    if alpha_delta <= tol and beta_delta <= tol:\n",
    "      break\n",
    "\n",
    "  return alpha, beta\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 173,
     "referenced_widgets": [
      "3bb9c178c2d84f5aab25c0023cdddc97",
      "34939a045991467291c684a105e35743",
      "a41e664d7ec34debbd5282e5a85dbd79",
      "7ce4487f9c974cc9ad359cfa54810f9c",
      "761b757ea7fe4a878380bf2fd4eefd8d",
      "0f399381e4bb402cb39ed03ce133fbbe",
      "f3574993fc2c41f39d70d8d99e85cfe3",
      "724bd045360d4c81978db55d0e6dc032"
     ]
    },
    "id": "HiTx3gyMOeEF",
    "outputId": "98ee9cdc-0263-458b-9670-0339b62e5602"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3bb9c178c2d84f5aab25c0023cdddc97",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0), HTML(value='')))"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/numpy/core/fromnumeric.py:3373: RuntimeWarning: Mean of empty slice.\n",
      "  out=out, **kwargs)\n",
      "/usr/local/lib/python3.6/dist-packages/numpy/core/_methods.py:170: RuntimeWarning: invalid value encountered in longdouble_scalars\n",
      "  ret = ret.dtype.type(ret / rcount)\n",
      "/usr/local/lib/python3.6/dist-packages/numpy/core/_methods.py:170: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  ret = ret.dtype.type(ret / rcount)\n"
     ]
    }
   ],
   "source": [
    "alpha, beta = m_step(docsTransformed, alpha0, beta0, phi0, gamma0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 435
    },
    "id": "w3Ai2N8iZLi_",
    "outputId": "047e1291-1ac7-4d10-f3b5-715ac1c3ae1c"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "      <th>34</th>\n",
       "      <th>35</th>\n",
       "      <th>36</th>\n",
       "      <th>37</th>\n",
       "      <th>38</th>\n",
       "      <th>39</th>\n",
       "      <th>...</th>\n",
       "      <th>960</th>\n",
       "      <th>961</th>\n",
       "      <th>962</th>\n",
       "      <th>963</th>\n",
       "      <th>964</th>\n",
       "      <th>965</th>\n",
       "      <th>966</th>\n",
       "      <th>967</th>\n",
       "      <th>968</th>\n",
       "      <th>969</th>\n",
       "      <th>970</th>\n",
       "      <th>971</th>\n",
       "      <th>972</th>\n",
       "      <th>973</th>\n",
       "      <th>974</th>\n",
       "      <th>975</th>\n",
       "      <th>976</th>\n",
       "      <th>977</th>\n",
       "      <th>978</th>\n",
       "      <th>979</th>\n",
       "      <th>980</th>\n",
       "      <th>981</th>\n",
       "      <th>982</th>\n",
       "      <th>983</th>\n",
       "      <th>984</th>\n",
       "      <th>985</th>\n",
       "      <th>986</th>\n",
       "      <th>987</th>\n",
       "      <th>988</th>\n",
       "      <th>989</th>\n",
       "      <th>990</th>\n",
       "      <th>991</th>\n",
       "      <th>992</th>\n",
       "      <th>993</th>\n",
       "      <th>994</th>\n",
       "      <th>995</th>\n",
       "      <th>996</th>\n",
       "      <th>997</th>\n",
       "      <th>998</th>\n",
       "      <th>999</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>880</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1327</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1361</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1399</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1580</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2699</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7250</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4216</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4511</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6519</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7930 rows × 1000 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      0    1    2    3    4    5    6    ...  993  994  995  996  997  998  999\n",
       "880   1.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "1327  1.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "1361  1.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "1399  1.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "1580  1.0  0.0  1.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  0.0  0.0  0.0  0.0\n",
       "...   ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...\n",
       "2699  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  1.0  0.0  0.0  0.0  0.0\n",
       "7250  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  1.0  0.0  0.0  0.0  0.0\n",
       "4216  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  1.0  0.0  0.0  0.0\n",
       "4511  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  6.0  0.0  0.0  0.0\n",
       "6519  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0  0.0  0.0  1.0  0.0\n",
       "\n",
       "[7930 rows x 1000 columns]"
      ]
     },
     "execution_count": 47,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "lk5HOJ5DXZnD"
   },
   "outputs": [],
   "source": [
    "def get_topic(ind):\n",
    "  topics = []\n",
    "  for i in docs[ind]:\n",
    "    ind = np.where(df.index==i)[0][0]\n",
    "    topicIdx = np.argmax(beta.T[ind])\n",
    "    # print(i, ind, topicIdx)\n",
    "    topics.append(topicIdx)\n",
    "  count = [topics.count(i) for i in topics]\n",
    "  try:\n",
    "    topic = np.argmax(count)\n",
    "  except:\n",
    "    topic = 0\n",
    "  return topic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "0xGtuG84Xbwq"
   },
   "outputs": [],
   "source": [
    "predLabels = []\n",
    "for ind, _ in enumerate(docs[:1000]):\n",
    "  predLabels.append(get_topic(ind))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "K-CU52WpetkX"
   },
   "outputs": [],
   "source": [
    "with open('/content/drive/MyDrive/Colab Notebooks/predLabels_VarInf_custom.pkl', 'wb') as f:\n",
    "  pickle.dump(predLabels, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "PLXiKivEfPIY"
   },
   "outputs": [],
   "source": [
    "file = open('/content/drive/MyDrive/Colab Notebooks/predLabels_VarInf_custom.pkl','rb')\n",
    "predLabels = pickle.load(file)\n",
    "file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 331
    },
    "id": "hq41ltKvdMho",
    "outputId": "e088a4b9-2fe0-4119-df73-bed6dd7e2f30"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "clusterNames = [\"Cluster%d\" % i for i in range(20)]\n",
    "targetNames = [\"Target%d\" % i for i in range(20)]\n",
    "cm = metrics.confusion_matrix(target[:1000], predLabels)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(plt, cm[:20, :20],\n",
    "                       classes=targetNames,\n",
    "                       clusters=clusterNames,\n",
    "                       title='LDA with custom Variational Inference')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ZlsP2ADMd_wH",
    "outputId": "b76ca3c9-feee-43d1-fb75-392025557874"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LDA with custom Variational Inference\n",
      "Homogeneity 0.06386051774632591\n",
      "Completenesee 0.08851532148057385\n",
      "Normalized Mutual Information 0.0741933142014031\n",
      "Purity 0.106\n"
     ]
    }
   ],
   "source": [
    "print(\"LDA with custom Variational Inference\")\n",
    "print(\"Homogeneity\", metrics.homogeneity_score(target[:1000], predLabels))\n",
    "print(\"Completenesee\", metrics.completeness_score(target[:1000], predLabels))\n",
    "print(\"Normalized Mutual Information\", metrics.normalized_mutual_info_score(target[:1000], predLabels))\n",
    "print(\"Purity\", purity_score(target[:1000], predLabels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "1-uQwfy9r_jT"
   },
   "source": [
    "# Comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 401
    },
    "id": "Lzk7EnrzsBUf",
    "outputId": "c7876cbb-e722-4ef3-8a00-7fa4ae94566e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity 0.04063416533673405\n",
      "Completeness 0.11067726730802309\n",
      "Normalized Mutual Information 0.05944399970583384\n",
      "Purity 0.09395439278769666\n"
     ]
    }
   ],
   "source": [
    "file = open('/content/drive/MyDrive/Colab Notebooks/predLabels_Gibbs_custom.pkl','rb')\n",
    "predLabels = pickle.load(file)\n",
    "file.close()\n",
    "\n",
    "clusterNames = [\"Cluster%d\" % i for i in range(20)]\n",
    "targetNames = [\"Target%d\" % i for i in range(20)]\n",
    "cm = metrics.confusion_matrix(target, predLabels)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(plt, cm,\n",
    "                       classes=targetNames,\n",
    "                       clusters=clusterNames,\n",
    "                       title='LDA with custom Gibbs sampling')\n",
    "plt.show()\n",
    "\n",
    "print(\"Homogeneity\", metrics.homogeneity_score(target, predLabels))\n",
    "print(\"Completeness\", metrics.completeness_score(target, predLabels))\n",
    "print(\"Normalized Mutual Information\", metrics.normalized_mutual_info_score(target, predLabels))\n",
    "print(\"Purity\", purity_score(target, predLabels))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 401
    },
    "id": "jg3ATO_JsPC9",
    "outputId": "888305a3-045c-4e3e-df55-875fa07ae064"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity 0.06386051774632591\n",
      "Completeness 0.08851532148057385\n",
      "Normalized Mutual Information 0.0741933142014031\n",
      "Purity 0.106\n"
     ]
    }
   ],
   "source": [
    "file = open('/content/drive/MyDrive/Colab Notebooks/predLabels_VarInf_custom.pkl','rb')\n",
    "predLabels = pickle.load(file)\n",
    "file.close()\n",
    "\n",
    "clusterNames = [\"Cluster%d\" % i for i in range(20)]\n",
    "targetNames = [\"Target%d\" % i for i in range(20)]\n",
    "cm = metrics.confusion_matrix(target[:1000], predLabels)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(plt, cm[:20, :20],\n",
    "                       classes=targetNames,\n",
    "                       clusters=clusterNames,\n",
    "                       title='LDA with custom Gibbs sampling')\n",
    "plt.show()\n",
    "\n",
    "print(\"Homogeneity\", metrics.homogeneity_score(target[:1000], predLabels))\n",
    "print(\"Completeness\", metrics.completeness_score(target[:1000], predLabels))\n",
    "print(\"Normalized Mutual Information\", metrics.normalized_mutual_info_score(target[:1000], predLabels))\n",
    "print(\"Purity\", purity_score(target[:1000], predLabels))"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "lda_custom_functions.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.7.0"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "00d59c369ac443d89786ef4b8040ac7a": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "0e34c7826bc34425ac785df9d50b0fab": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "0f399381e4bb402cb39ed03ce133fbbe": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "1a00528558194b6b893ba88548263173": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "1fbe6bef23184eecb7efb24036196bdb": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "100%",
      "description_tooltip": null,
      "layout": "IPY_MODEL_6e9f1e99023b4224b83e98dea1fe84f3",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_618c789eccd1447db7eaab21aae7da9b",
      "value": 1
     }
    },
    "227be25701ec4e92b20f475a858b854f": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "3277a95c4b1d4a278fae78f99f5b42eb": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_00d59c369ac443d89786ef4b8040ac7a",
      "placeholder": "​",
      "style": "IPY_MODEL_4b9af6a945ee47f982b362702fd60327",
      "value": " 1/1 [42:39&lt;00:00, 2559.53s/it]"
     }
    },
    "34939a045991467291c684a105e35743": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "3bb9c178c2d84f5aab25c0023cdddc97": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_a41e664d7ec34debbd5282e5a85dbd79",
       "IPY_MODEL_7ce4487f9c974cc9ad359cfa54810f9c"
      ],
      "layout": "IPY_MODEL_34939a045991467291c684a105e35743"
     }
    },
    "4b9af6a945ee47f982b362702fd60327": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "618c789eccd1447db7eaab21aae7da9b": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": "initial"
     }
    },
    "6885b492588342afb1403a7c0bce99af": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "100%",
      "description_tooltip": null,
      "layout": "IPY_MODEL_1a00528558194b6b893ba88548263173",
      "max": 10,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_7dbbf0e6e62d4ae3875b20dfee860d14",
      "value": 10
     }
    },
    "69a349e5087142ad8132e303e499b9d6": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_6885b492588342afb1403a7c0bce99af",
       "IPY_MODEL_8da1c14f46ab4e0cb1d166636d65ca7d"
      ],
      "layout": "IPY_MODEL_e99b07f70b8b4740abfc29b53cb99617"
     }
    },
    "6e9f1e99023b4224b83e98dea1fe84f3": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "724bd045360d4c81978db55d0e6dc032": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "761b757ea7fe4a878380bf2fd4eefd8d": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": "initial"
     }
    },
    "7ce4487f9c974cc9ad359cfa54810f9c": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_724bd045360d4c81978db55d0e6dc032",
      "placeholder": "​",
      "style": "IPY_MODEL_f3574993fc2c41f39d70d8d99e85cfe3",
      "value": " 35/100 [08:31&lt;09:16,  8.56s/it]"
     }
    },
    "7dbbf0e6e62d4ae3875b20dfee860d14": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": "initial"
     }
    },
    "87aa2524714942639f2317d2f11b4d9f": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "8da1c14f46ab4e0cb1d166636d65ca7d": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_0e34c7826bc34425ac785df9d50b0fab",
      "placeholder": "​",
      "style": "IPY_MODEL_87aa2524714942639f2317d2f11b4d9f",
      "value": " 10/10 [42:35&lt;00:00, 255.52s/it]"
     }
    },
    "a41e664d7ec34debbd5282e5a85dbd79": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "danger",
      "description": " 35%",
      "description_tooltip": null,
      "layout": "IPY_MODEL_0f399381e4bb402cb39ed03ce133fbbe",
      "max": 100,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_761b757ea7fe4a878380bf2fd4eefd8d",
      "value": 35
     }
    },
    "d39aa5ddaafa4c7a8efca9a7123357bc": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_1fbe6bef23184eecb7efb24036196bdb",
       "IPY_MODEL_3277a95c4b1d4a278fae78f99f5b42eb"
      ],
      "layout": "IPY_MODEL_227be25701ec4e92b20f475a858b854f"
     }
    },
    "e99b07f70b8b4740abfc29b53cb99617": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f3574993fc2c41f39d70d8d99e85cfe3": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}