Format: application/octet-stream
⬇ Download File (55.2 KB)
   1  {
   2   "cells": [
   3    {
   4     "cell_type": "code",
   5     "execution_count": 1,
   6     "id": "954e9fd9",
   7     "metadata": {},
   8     "outputs": [],
   9     "source": [
  10      "import pandas as pd\n",
  11      "import numpy as np\n",
  12      "import nltk\n",
  13      "from nltk.corpus import stopwords\n",
  14      "stopwords= stopwords.words('english')\n",
  15      "stopwords.extend([',','`','?',\"'s\", \"''\", '``',\"n't\",\"(\", \")\", ';',\"'\",'re','!','--','-','*',\"'re\",':'])"
  16     ]
  17    },
  18    {
  19     "cell_type": "code",
  20     "execution_count": 2,
  21     "id": "5c9f32c2",
  22     "metadata": {},
  23     "outputs": [],
  24     "source": [
  25      "stopwords.extend([i.capitalize() for i in stopwords])"
  26     ]
  27    },
  28    {
  29     "cell_type": "code",
  30     "execution_count": 3,
  31     "id": "18a56afa",
  32     "metadata": {},
  33     "outputs": [],
  34     "source": [
  35      "s1= pd.read_csv('data/unique synopsis.tsv', sep='\\t')\n"
  36     ]
  37    },
  38    {
  39     "cell_type": "code",
  40     "execution_count": 4,
  41     "id": "f07f0c11",
  42     "metadata": {
  43      "scrolled": false
  44     },
  45     "outputs": [
  46      {
  47       "data": {
  48        "text/html": [
  49         "<div>\n",
  50         "<style scoped>\n",
  51         "    .dataframe tbody tr th:only-of-type {\n",
  52         "        vertical-align: middle;\n",
  53         "    }\n",
  54         "\n",
  55         "    .dataframe tbody tr th {\n",
  56         "        vertical-align: top;\n",
  57         "    }\n",
  58         "\n",
  59         "    .dataframe thead th {\n",
  60         "        text-align: right;\n",
  61         "    }\n",
  62         "</style>\n",
  63         "<table border=\"1\" class=\"dataframe\">\n",
  64         "  <thead>\n",
  65         "    <tr style=\"text-align: right;\">\n",
  66         "      <th></th>\n",
  67         "      <th>0</th>\n",
  68         "    </tr>\n",
  69         "  </thead>\n",
  70         "  <tbody>\n",
  71         "    <tr>\n",
  72         "      <th>0</th>\n",
  73         "      <td>In 1947, Andy Dufresne (Tim Robbins), a banker...</td>\n",
  74         "    </tr>\n",
  75         "    <tr>\n",
  76         "      <th>1</th>\n",
  77         "      <td>The movie begins with a gang of men with clown...</td>\n",
  78         "    </tr>\n",
  79         "    <tr>\n",
  80         "      <th>2</th>\n",
  81         "      <td>A young man, exhausted and delirious, washes u...</td>\n",
  82         "    </tr>\n",
  83         "    <tr>\n",
  84         "      <th>3</th>\n",
  85         "      <td>We back out of the webbing of neurons and brai...</td>\n",
  86         "    </tr>\n",
  87         "    <tr>\n",
  88         "      <th>4</th>\n",
  89         "      <td>Late one morning in the Hawthorne Grill, a res...</td>\n",
  90         "    </tr>\n",
  91         "  </tbody>\n",
  92         "</table>\n",
  93         "</div>"
  94        ],
  95        "text/plain": [
  96         "                                                   0\n",
  97         "0  In 1947, Andy Dufresne (Tim Robbins), a banker...\n",
  98         "1  The movie begins with a gang of men with clown...\n",
  99         "2  A young man, exhausted and delirious, washes u...\n",
 100         "3  We back out of the webbing of neurons and brai...\n",
 101         "4  Late one morning in the Hawthorne Grill, a res..."
 102        ]
 103       },
 104       "execution_count": 4,
 105       "metadata": {},
 106       "output_type": "execute_result"
 107      }
 108     ],
 109     "source": [
 110      "s1.head()"
 111     ]
 112    },
 113    {
 114     "cell_type": "code",
 115     "execution_count": 5,
 116     "id": "a52e3b31",
 117     "metadata": {},
 118     "outputs": [
 119      {
 120       "data": {
 121        "text/plain": [
 122         "9643"
 123        ]
 124       },
 125       "execution_count": 5,
 126       "metadata": {},
 127       "output_type": "execute_result"
 128      }
 129     ],
 130     "source": [
 131      "len(s1)"
 132     ]
 133    },
 134    {
 135     "cell_type": "code",
 136     "execution_count": 6,
 137     "id": "74d6ffd7",
 138     "metadata": {},
 139     "outputs": [],
 140     "source": [
 141      "from nltk.tokenize import word_tokenize"
 142     ]
 143    },
 144    {
 145     "cell_type": "code",
 146     "execution_count": 7,
 147     "id": "50095664",
 148     "metadata": {},
 149     "outputs": [],
 150     "source": [
 151      "texts= s1.dropna(axis=0).values.reshape(-1,).astype('object')"
 152     ]
 153    },
 154    {
 155     "cell_type": "code",
 156     "execution_count": 8,
 157     "id": "d1ea9fdd",
 158     "metadata": {},
 159     "outputs": [],
 160     "source": [
 161      "texts=' '.join(texts)"
 162     ]
 163    },
 164    {
 165     "cell_type": "code",
 166     "execution_count": 9,
 167     "id": "10634606",
 168     "metadata": {},
 169     "outputs": [],
 170     "source": [
 171      "texts=' '.join(texts.split('.')).lower()"
 172     ]
 173    },
 174    {
 175     "cell_type": "code",
 176     "execution_count": null,
 177     "id": "3c3adcaf",
 178     "metadata": {},
 179     "outputs": [],
 180     "source": []
 181    },
 182    {
 183     "cell_type": "code",
 184     "execution_count": 10,
 185     "id": "cd909ad7",
 186     "metadata": {},
 187     "outputs": [],
 188     "source": [
 189      "words= np.array(word_tokenize(texts))  ### chewck"
 190     ]
 191    },
 192    {
 193     "cell_type": "code",
 194     "execution_count": 11,
 195     "id": "661ed7bd",
 196     "metadata": {},
 197     "outputs": [],
 198     "source": [
 199      "words=np.array([w for w in words if not w in stopwords])"
 200     ]
 201    },
 202    {
 203     "cell_type": "code",
 204     "execution_count": 12,
 205     "id": "2fe22ff2",
 206     "metadata": {},
 207     "outputs": [
 208      {
 209       "data": {
 210        "text/plain": [
 211         "5641411"
 212        ]
 213       },
 214       "execution_count": 12,
 215       "metadata": {},
 216       "output_type": "execute_result"
 217      }
 218     ],
 219     "source": [
 220      "len(words)"
 221     ]
 222    },
 223    {
 224     "cell_type": "code",
 225     "execution_count": 13,
 226     "id": "62986f26",
 227     "metadata": {},
 228     "outputs": [],
 229     "source": [
 230      "features=np.array(list(set(words)))"
 231     ]
 232    },
 233    {
 234     "cell_type": "code",
 235     "execution_count": 14,
 236     "id": "0373e926",
 237     "metadata": {},
 238     "outputs": [
 239      {
 240       "data": {
 241        "text/plain": [
 242         "128962"
 243        ]
 244       },
 245       "execution_count": 14,
 246       "metadata": {},
 247       "output_type": "execute_result"
 248      }
 249     ],
 250     "source": [
 251      "len(features)"
 252     ]
 253    },
 254    {
 255     "cell_type": "code",
 256     "execution_count": null,
 257     "id": "c0f901b4",
 258     "metadata": {},
 259     "outputs": [],
 260     "source": []
 261    },
 262    {
 263     "cell_type": "code",
 264     "execution_count": 15,
 265     "id": "ddc5ac8d",
 266     "metadata": {},
 267     "outputs": [
 268      {
 269       "data": {
 270        "text/plain": [
 271         "array([], dtype=int64)"
 272        ]
 273       },
 274       "execution_count": 15,
 275       "metadata": {},
 276       "output_type": "execute_result"
 277      }
 278     ],
 279     "source": [
 280      "np.where(words== 'in')[0]"
 281     ]
 282    },
 283    {
 284     "cell_type": "code",
 285     "execution_count": 16,
 286     "id": "35406a83",
 287     "metadata": {},
 288     "outputs": [],
 289     "source": [
 290      "def count_f(array):\n",
 291      "    unique, counts = np.unique(array, return_counts=True)\n",
 292      "    if len(counts)>1:\n",
 293      "        counts= counts/len(counts)\n",
 294      "    else:\n",
 295      "        counts= counts/counts\n",
 296      "    array=dict(zip(unique, counts))\n",
 297      "    array={k: v for k, v in sorted(array.items(), key=lambda item: item[1], reverse=True)} #set to true\n",
 298      "    #print(array)\n",
 299      "    return array"
 300     ]
 301    },
 302    {
 303     "cell_type": "code",
 304     "execution_count": 17,
 305     "id": "d7a75ae7",
 306     "metadata": {},
 307     "outputs": [],
 308     "source": [
 309      "def next_previous(feature):\n",
 310      "    indexs= np.where(words== feature)[0]\n",
 311      "    \n",
 312      "    next_word= count_f(words[indexs+1])\n",
 313      "    previous_word= count_f(words[indexs-1])\n",
 314      "\n",
 315      "    return np.vstack((previous_word, next_word)).astype('object').T"
 316     ]
 317    },
 318    {
 319     "cell_type": "code",
 320     "execution_count": null,
 321     "id": "38253854",
 322     "metadata": {},
 323     "outputs": [],
 324     "source": []
 325    },
 326    {
 327     "cell_type": "code",
 328     "execution_count": null,
 329     "id": "977a0bc0",
 330     "metadata": {},
 331     "outputs": [],
 332     "source": []
 333    },
 334    {
 335     "cell_type": "code",
 336     "execution_count": null,
 337     "id": "3acfa8e6",
 338     "metadata": {},
 339     "outputs": [],
 340     "source": []
 341    },
 342    {
 343     "cell_type": "code",
 344     "execution_count": 18,
 345     "id": "7bdf69e7",
 346     "metadata": {},
 347     "outputs": [],
 348     "source": [
 349      "def word_cha(features):\n",
 350      "    final={}\n",
 351      "    for feature in features:\n",
 352      "        final[feature]= next_previous(feature)\n",
 353      "    return final"
 354     ]
 355    },
 356    {
 357     "cell_type": "code",
 358     "execution_count": 27,
 359     "id": "db74d8c2",
 360     "metadata": {},
 361     "outputs": [],
 362     "source": [
 363      "def find_most_probable(word,num=10):\n",
 364      "    try:\n",
 365      "        words=next_previous(word)\n",
 366      "        previous_words= np.array(list(words[0][0].keys())).astype('object')[:num]\n",
 367      "        next_words= np.array(list(words[0][1].keys())).astype('object')[:num]\n",
 368      "        previous_words_prob= np.array(list(words[0][0].values()))\n",
 369      "        prob_words=np.array([])\n",
 370      "        for i,previous_word in enumerate(previous_words):\n",
 371      "            #print(i, 'p',previous_word)\n",
 372      "            #c=previous_words_prob[i]\n",
 373      "            #print(f'c= {c}')\n",
 374      "            n=next_previous(previous_word)\n",
 375      "            #print(n)\n",
 376      "\n",
 377      "            next_words_child= np.array(list(n[0][1].keys())).astype('object')\n",
 378      "            prob_words= np.append(prob_words, next_words_child)\n",
 379      "\n",
 380      "            if i<len(next_words):\n",
 381      "                n=next_previous(next_words[i])\n",
 382      "                previous_words_child= np.array(list(n[0][0].keys())).astype('object')\n",
 383      "                prob_words= np.append(prob_words, previous_words_child)\n",
 384      "                #print(next_words_child)\n",
 385      "                #next_words_child_prob= np.array(list(next_previous(word)[0][1].values()))\n",
 386      "\n",
 387      "     \n",
 388      "        #del prob_words[word]\n",
 389      "        #a=count_f(words)\n",
 390      "        #for key, value in prob_words.items():\n",
 391      "        #    prob_words[key]= value/a[key]\n",
 392      "       # prob_words={k: v for k, v in sorted(prob_words.items(), key=lambda item: item[1], reverse=True)}\n",
 393      "    except:\n",
 394      "        pass\n",
 395      "    prob_words=count_f(prob_words)\n",
 396      "    #del prob_words[word]\n",
 397      "    return prob_words"
 398     ]
 399    },
 400    {
 401     "cell_type": "code",
 402     "execution_count": null,
 403     "id": "735f09df",
 404     "metadata": {},
 405     "outputs": [],
 406     "source": []
 407    },
 408    {
 409     "cell_type": "code",
 410     "execution_count": 28,
 411     "id": "0253d5f9",
 412     "metadata": {},
 413     "outputs": [
 414      {
 415       "data": {
 416        "text/plain": [
 417         "{'leaves': 0.000765579543714592,\n",
 418         " 'make': 0.000765579543714592,\n",
 419         " 'men': 0.000765579543714592,\n",
 420         " 'money': 0.000765579543714592,\n",
 421         " 'one': 0.000765579543714592,\n",
 422         " 'people': 0.000765579543714592,\n",
 423         " 'time': 0.000765579543714592,\n",
 424         " 'another': 0.0007273005665288624,\n",
 425         " 'asks': 0.0007273005665288624,\n",
 426         " 'back': 0.0007273005665288624,\n",
 427         " 'car': 0.0007273005665288624,\n",
 428         " 'cash': 0.0007273005665288624,\n",
 429         " 'day': 0.0007273005665288624,\n",
 430         " 'drugs': 0.0007273005665288624,\n",
 431         " 'even': 0.0007273005665288624,\n",
 432         " 'everything': 0.0007273005665288624,\n",
 433         " 'father': 0.0007273005665288624,\n",
 434         " 'fight': 0.0007273005665288624,\n",
 435         " 'first': 0.0007273005665288624,\n",
 436         " 'food': 0.0007273005665288624,\n",
 437         " 'girls': 0.0007273005665288624,\n",
 438         " 'go': 0.0007273005665288624,\n",
 439         " 'jack': 0.0007273005665288624,\n",
 440         " 'man': 0.0007273005665288624,\n",
 441         " 'needs': 0.0007273005665288624,\n",
 442         " 'right': 0.0007273005665288624,\n",
 443         " 'says': 0.0007273005665288624,\n",
 444         " 'tells': 0.0007273005665288624,\n",
 445         " 'things': 0.0007273005665288624,\n",
 446         " 'two': 0.0007273005665288624,\n",
 447         " 'andy': 0.0006890215893431327,\n",
 448         " 'bad': 0.0006890215893431327,\n",
 449         " 'blood': 0.0006890215893431327,\n",
 450         " 'charlie': 0.0006890215893431327,\n",
 451         " 'danny': 0.0006890215893431327,\n",
 452         " 'david': 0.0006890215893431327,\n",
 453         " 'dead': 0.0006890215893431327,\n",
 454         " 'everyone': 0.0006890215893431327,\n",
 455         " 'family': 0.0006890215893431327,\n",
 456         " 'frank': 0.0006890215893431327,\n",
 457         " 'free': 0.0006890215893431327,\n",
 458         " 'friends': 0.0006890215893431327,\n",
 459         " 'george': 0.0006890215893431327,\n",
 460         " 'gets': 0.0006890215893431327,\n",
 461         " 'girl': 0.0006890215893431327,\n",
 462         " 'goes': 0.0006890215893431327,\n",
 463         " 'good': 0.0006890215893431327,\n",
 464         " 'group': 0.0006890215893431327,\n",
 465         " 'head': 0.0006890215893431327,\n",
 466         " 'instead': 0.0006890215893431327,\n",
 467         " 'jim': 0.0006890215893431327,\n",
 468         " 'john': 0.0006890215893431327,\n",
 469         " 'kids': 0.0006890215893431327,\n",
 470         " 'like': 0.0006890215893431327,\n",
 471         " 'live': 0.0006890215893431327,\n",
 472         " 'long': 0.0006890215893431327,\n",
 473         " 'mother': 0.0006890215893431327,\n",
 474         " 'movie': 0.0006890215893431327,\n",
 475         " 'much': 0.0006890215893431327,\n",
 476         " 'next': 0.0006890215893431327,\n",
 477         " 'parents': 0.0006890215893431327,\n",
 478         " 'part': 0.0006890215893431327,\n",
 479         " 'room': 0.0006890215893431327,\n",
 480         " 'runs': 0.0006890215893431327,\n",
 481         " 'sarah': 0.0006890215893431327,\n",
 482         " 'show': 0.0006890215893431327,\n",
 483         " 'since': 0.0006890215893431327,\n",
 484         " 'something': 0.0006890215893431327,\n",
 485         " 'soon': 0.0006890215893431327,\n",
 486         " 'stop': 0.0006890215893431327,\n",
 487         " 'story': 0.0006890215893431327,\n",
 488         " 'takes': 0.0006890215893431327,\n",
 489         " 'three': 0.0006890215893431327,\n",
 490         " 'tony': 0.0006890215893431327,\n",
 491         " 'want': 0.0006890215893431327,\n",
 492         " 'water': 0.0006890215893431327,\n",
 493         " 'wife': 0.0006890215893431327,\n",
 494         " 'without': 0.0006890215893431327,\n",
 495         " 'alex': 0.0006507426121574032,\n",
 496         " 'anna': 0.0006507426121574032,\n",
 497         " 'away': 0.0006507426121574032,\n",
 498         " 'ben': 0.0006507426121574032,\n",
 499         " 'better': 0.0006507426121574032,\n",
 500         " 'board': 0.0006507426121574032,\n",
 501         " 'body': 0.0006507426121574032,\n",
 502         " 'boys': 0.0006507426121574032,\n",
 503         " 'business': 0.0006507426121574032,\n",
 504         " 'call': 0.0006507426121574032,\n",
 505         " 'check': 0.0006507426121574032,\n",
 506         " 'children': 0.0006507426121574032,\n",
 507         " 'claire': 0.0006507426121574032,\n",
 508         " 'come': 0.0006507426121574032,\n",
 509         " 'company': 0.0006507426121574032,\n",
 510         " 'could': 0.0006507426121574032,\n",
 511         " 'couple': 0.0006507426121574032,\n",
 512         " 'cut': 0.0006507426121574032,\n",
 513         " 'daughter': 0.0006507426121574032,\n",
 514         " 'film': 0.0006507426121574032,\n",
 515         " 'finally': 0.0006507426121574032,\n",
 516         " 'find': 0.0006507426121574032,\n",
 517         " 'fish': 0.0006507426121574032,\n",
 518         " 'get': 0.0006507426121574032,\n",
 519         " 'harry': 0.0006507426121574032,\n",
 520         " 'helen': 0.0006507426121574032,\n",
 521         " 'help': 0.0006507426121574032,\n",
 522         " 'henry': 0.0006507426121574032,\n",
 523         " 'house': 0.0006507426121574032,\n",
 524         " 'however': 0.0006507426121574032,\n",
 525         " 'information': 0.0006507426121574032,\n",
 526         " 'job': 0.0006507426121574032,\n",
 527         " 'joe': 0.0006507426121574032,\n",
 528         " 'johnny': 0.0006507426121574032,\n",
 529         " 'life': 0.0006507426121574032,\n",
 530         " 'looking': 0.0006507426121574032,\n",
 531         " 'love': 0.0006507426121574032,\n",
 532         " 'maggie': 0.0006507426121574032,\n",
 533         " 'mike': 0.0006507426121574032,\n",
 534         " 'night': 0.0006507426121574032,\n",
 535         " 'nothing': 0.0006507426121574032,\n",
 536         " 'order': 0.0006507426121574032,\n",
 537         " 'outside': 0.0006507426121574032,\n",
 538         " 'paul': 0.0006507426121574032,\n",
 539         " 'person': 0.0006507426121574032,\n",
 540         " 'peter': 0.0006507426121574032,\n",
 541         " 'place': 0.0006507426121574032,\n",
 542         " 'plan': 0.0006507426121574032,\n",
 543         " 'red': 0.0006507426121574032,\n",
 544         " 'sam': 0.0006507426121574032,\n",
 545         " 'school': 0.0006507426121574032,\n",
 546         " 'security': 0.0006507426121574032,\n",
 547         " 'see': 0.0006507426121574032,\n",
 548         " 'set': 0.0006507426121574032,\n",
 549         " 'son': 0.0006507426121574032,\n",
 550         " 'tom': 0.0006507426121574032,\n",
 551         " 'wants': 0.0006507426121574032,\n",
 552         " 'way': 0.0006507426121574032,\n",
 553         " 'white': 0.0006507426121574032,\n",
 554         " 'woman': 0.0006507426121574032,\n",
 555         " 'work': 0.0006507426121574032,\n",
 556         " 'would': 0.0006507426121574032,\n",
 557         " 'adam': 0.0006124636349716736,\n",
 558         " 'alice': 0.0006124636349716736,\n",
 559         " 'also': 0.0006124636349716736,\n",
 560         " 'answer': 0.0006124636349716736,\n",
 561         " 'anything': 0.0006124636349716736,\n",
 562         " 'apartment': 0.0006124636349716736,\n",
 563         " 'attempt': 0.0006124636349716736,\n",
 564         " 'baby': 0.0006124636349716736,\n",
 565         " 'big': 0.0006124636349716736,\n",
 566         " 'black': 0.0006124636349716736,\n",
 567         " 'boat': 0.0006124636349716736,\n",
 568         " 'bobby': 0.0006124636349716736,\n",
 569         " 'bond': 0.0006124636349716736,\n",
 570         " 'box': 0.0006124636349716736,\n",
 571         " 'boy': 0.0006124636349716736,\n",
 572         " 'brother': 0.0006124636349716736,\n",
 573         " 'bruce': 0.0006124636349716736,\n",
 574         " 'building': 0.0006124636349716736,\n",
 575         " 'change': 0.0006124636349716736,\n",
 576         " 'chris': 0.0006124636349716736,\n",
 577         " 'church': 0.0006124636349716736,\n",
 578         " 'control': 0.0006124636349716736,\n",
 579         " 'cover': 0.0006124636349716736,\n",
 580         " 'dad': 0.0006124636349716736,\n",
 581         " 'daniel': 0.0006124636349716736,\n",
 582         " 'deal': 0.0006124636349716736,\n",
 583         " 'death': 0.0006124636349716736,\n",
 584         " 'dinner': 0.0006124636349716736,\n",
 585         " 'drink': 0.0006124636349716736,\n",
 586         " 'drive': 0.0006124636349716736,\n",
 587         " 'far': 0.0006124636349716736,\n",
 588         " 'game': 0.0006124636349716736,\n",
 589         " 'give': 0.0006124636349716736,\n",
 590         " 'going': 0.0006124636349716736,\n",
 591         " 'gold': 0.0006124636349716736,\n",
 592         " 'ground': 0.0006124636349716736,\n",
 593         " 'hand': 0.0006124636349716736,\n",
 594         " 'hard': 0.0006124636349716736,\n",
 595         " 'heart': 0.0006124636349716736,\n",
 596         " 'home': 0.0006124636349716736,\n",
 597         " 'idea': 0.0006124636349716736,\n",
 598         " 'inside': 0.0006124636349716736,\n",
 599         " 'jacob': 0.0006124636349716736,\n",
 600         " 'jane': 0.0006124636349716736,\n",
 601         " 'jesse': 0.0006124636349716736,\n",
 602         " 'julie': 0.0006124636349716736,\n",
 603         " 'junior': 0.0006124636349716736,\n",
 604         " 'kate': 0.0006124636349716736,\n",
 605         " 'kevin': 0.0006124636349716736,\n",
 606         " 'killing': 0.0006124636349716736,\n",
 607         " 'last': 0.0006124636349716736,\n",
 608         " 'later': 0.0006124636349716736,\n",
 609         " 'laura': 0.0006124636349716736,\n",
 610         " 'leave': 0.0006124636349716736,\n",
 611         " 'leaving': 0.0006124636349716736,\n",
 612         " 'left': 0.0006124636349716736,\n",
 613         " 'look': 0.0006124636349716736,\n",
 614         " 'lot': 0.0006124636349716736,\n",
 615         " 'mark': 0.0006124636349716736,\n",
 616         " 'max': 0.0006124636349716736,\n",
 617         " 'michael': 0.0006124636349716736,\n",
 618         " 'never': 0.0006124636349716736,\n",
 619         " 'nick': 0.0006124636349716736,\n",
 620         " 'offer': 0.0006124636349716736,\n",
 621         " 'party': 0.0006124636349716736,\n",
 622         " 'pay': 0.0006124636349716736,\n",
 623         " 'police': 0.0006124636349716736,\n",
 624         " 'power': 0.0006124636349716736,\n",
 625         " 'real': 0.0006124636349716736,\n",
 626         " 'rest': 0.0006124636349716736,\n",
 627         " 'sees': 0.0006124636349716736,\n",
 628         " 'someone': 0.0006124636349716736,\n",
 629         " 'steve': 0.0006124636349716736,\n",
 630         " 'still': 0.0006124636349716736,\n",
 631         " 'talk': 0.0006124636349716736,\n",
 632         " 'team': 0.0006124636349716736,\n",
 633         " 'terry': 0.0006124636349716736,\n",
 634         " 'though': 0.0006124636349716736,\n",
 635         " 'town': 0.0006124636349716736,\n",
 636         " 'trouble': 0.0006124636349716736,\n",
 637         " 'try': 0.0006124636349716736,\n",
 638         " 'us': 0.0006124636349716736,\n",
 639         " 'use': 0.0006124636349716736,\n",
 640         " 'wedding': 0.0006124636349716736,\n",
 641         " 'well': 0.0006124636349716736,\n",
 642         " 'world': 0.0006124636349716736,\n",
 643         " 'wrong': 0.0006124636349716736,\n",
 644         " \"'ll\": 0.0005741846577859439,\n",
 645         " 'advice': 0.0005741846577859439,\n",
 646         " 'amy': 0.0005741846577859439,\n",
 647         " 'arthur': 0.0005741846577859439,\n",
 648         " 'attention': 0.0005741846577859439,\n",
 649         " 'bag': 0.0005741846577859439,\n",
 650         " 'beth': 0.0005741846577859439,\n",
 651         " 'billy': 0.0005741846577859439,\n",
 652         " 'bob': 0.0005741846577859439,\n",
 653         " 'break': 0.0005741846577859439,\n",
 654         " 'brian': 0.0005741846577859439,\n",
 655         " 'brothers': 0.0005741846577859439,\n",
 656         " 'bud': 0.0005741846577859439,\n",
 657         " 'bus': 0.0005741846577859439,\n",
 658         " 'ca': 0.0005741846577859439,\n",
 659         " 'calls': 0.0005741846577859439,\n",
 660         " 'card': 0.0005741846577859439,\n",
 661         " 'carl': 0.0005741846577859439,\n",
 662         " 'case': 0.0005741846577859439,\n",
 663         " 'cat': 0.0005741846577859439,\n",
 664         " 'catherine': 0.0005741846577859439,\n",
 665         " 'chance': 0.0005741846577859439,\n",
 666         " 'chase': 0.0005741846577859439,\n",
 667         " 'child': 0.0005741846577859439,\n",
 668         " 'city': 0.0005741846577859439,\n",
 669         " 'class': 0.0005741846577859439,\n",
 670         " 'clothes': 0.0005741846577859439,\n",
 671         " 'cold': 0.0005741846577859439,\n",
 672         " 'cops': 0.0005741846577859439,\n",
 673         " 'dance': 0.0005741846577859439,\n",
 674         " 'dave': 0.0005741846577859439,\n",
 675         " 'dean': 0.0005741846577859439,\n",
 676         " 'dog': 0.0005741846577859439,\n",
 677         " 'door': 0.0005741846577859439,\n",
 678         " 'driver': 0.0005741846577859439,\n",
 679         " 'eddie': 0.0005741846577859439,\n",
 680         " 'effort': 0.0005741846577859439,\n",
 681         " 'either': 0.0005741846577859439,\n",
 682         " 'emily': 0.0005741846577859439,\n",
 683         " 'enough': 0.0005741846577859439,\n",
 684         " 'escape': 0.0005741846577859439,\n",
 685         " 'eventually': 0.0005741846577859439,\n",
 686         " 'evidence': 0.0005741846577859439,\n",
 687         " 'finds': 0.0005741846577859439,\n",
 688         " 'form': 0.0005741846577859439,\n",
 689         " 'friend': 0.0005741846577859439,\n",
 690         " 'gang': 0.0005741846577859439,\n",
 691         " 'getting': 0.0005741846577859439,\n",
 692         " 'guard': 0.0005741846577859439,\n",
 693         " 'gun': 0.0005741846577859439,\n",
 694         " 'guy': 0.0005741846577859439,\n",
 695         " 'high': 0.0005741846577859439,\n",
 696         " 'hospital': 0.0005741846577859439,\n",
 697         " 'hotel': 0.0005741846577859439,\n",
 698         " 'husband': 0.0005741846577859439,\n",
 699         " 'jake': 0.0005741846577859439,\n",
 700         " 'karl': 0.0005741846577859439,\n",
 701         " 'keep': 0.0005741846577859439,\n",
 702         " 'kid': 0.0005741846577859439,\n",
 703         " 'killer': 0.0005741846577859439,\n",
 704         " 'know': 0.0005741846577859439,\n",
 705         " 'kyle': 0.0005741846577859439,\n",
 706         " 'light': 0.0005741846577859439,\n",
 707         " 'lisa': 0.0005741846577859439,\n",
 708         " 'little': 0.0005741846577859439,\n",
 709         " 'lives': 0.0005741846577859439,\n",
 710         " 'looks': 0.0005741846577859439,\n",
 711         " 'lou': 0.0005741846577859439,\n",
 712         " 'lucy': 0.0005741846577859439,\n",
 713         " 'marty': 0.0005741846577859439,\n",
 714         " 'mary': 0.0005741846577859439,\n",
 715         " 'matt': 0.0005741846577859439,\n",
 716         " 'meet': 0.0005741846577859439,\n",
 717         " 'message': 0.0005741846577859439,\n",
 718         " 'moment': 0.0005741846577859439,\n",
 719         " 'morning': 0.0005741846577859439,\n",
 720         " 'name': 0.0005741846577859439,\n",
 721         " 'need': 0.0005741846577859439,\n",
 722         " 'nina': 0.0005741846577859439,\n",
 723         " 'norman': 0.0005741846577859439,\n",
 724         " 'number': 0.0005741846577859439,\n",
 725         " 'others': 0.0005741846577859439,\n",
 726         " 'paper': 0.0005741846577859439,\n",
 727         " 'pete': 0.0005741846577859439,\n",
 728         " 'phone': 0.0005741846577859439,\n",
 729         " 'plans': 0.0005741846577859439,\n",
 730         " 'ray': 0.0005741846577859439,\n",
 731         " 'realizes': 0.0005741846577859439,\n",
 732         " 'really': 0.0005741846577859439,\n",
 733         " 'refuses': 0.0005741846577859439,\n",
 734         " 'relationship': 0.0005741846577859439,\n",
 735         " 'return': 0.0005741846577859439,\n",
 736         " 'rose': 0.0005741846577859439,\n",
 737         " 'sally': 0.0005741846577859439,\n",
 738         " 'sara': 0.0005741846577859439,\n",
 739         " 'say': 0.0005741846577859439,\n",
 740         " 'saying': 0.0005741846577859439,\n",
 741         " 'scene': 0.0005741846577859439,\n",
 742         " 'second': 0.0005741846577859439,\n",
 743         " 'sex': 0.0005741846577859439,\n",
 744         " 'shot': 0.0005741846577859439,\n",
 745         " 'sister': 0.0005741846577859439,\n",
 746         " 'soldiers': 0.0005741846577859439,\n",
 747         " 'start': 0.0005741846577859439,\n",
 748         " 'station': 0.0005741846577859439,\n",
 749         " 'stay': 0.0005741846577859439,\n",
 750         " 'store': 0.0005741846577859439,\n",
 751         " 'street': 0.0005741846577859439,\n",
 752         " 'take': 0.0005741846577859439,\n",
 753         " 'taking': 0.0005741846577859439,\n",
 754         " 'tell': 0.0005741846577859439,\n",
 755         " 'thought': 0.0005741846577859439,\n",
 756         " 'tries': 0.0005741846577859439,\n",
 757         " 'true': 0.0005741846577859439,\n",
 758         " 'voice': 0.0005741846577859439,\n",
 759         " 'watch': 0.0005741846577859439,\n",
 760         " 'weapons': 0.0005741846577859439,\n",
 761         " 'women': 0.0005741846577859439,\n",
 762         " 'years': 0.0005741846577859439,\n",
 763         " 'able': 0.0005359056806002144,\n",
 764         " 'agrees': 0.0005359056806002144,\n",
 765         " 'air': 0.0005359056806002144,\n",
 766         " 'airport': 0.0005359056806002144,\n",
 767         " 'alien': 0.0005359056806002144,\n",
 768         " 'aliens': 0.0005359056806002144,\n",
 769         " 'alive': 0.0005359056806002144,\n",
 770         " 'alone': 0.0005359056806002144,\n",
 771         " 'along': 0.0005359056806002144,\n",
 772         " 'anger': 0.0005359056806002144,\n",
 773         " 'annie': 0.0005359056806002144,\n",
 774         " 'answers': 0.0005359056806002144,\n",
 775         " 'anyone': 0.0005359056806002144,\n",
 776         " 'area': 0.0005359056806002144,\n",
 777         " 'army': 0.0005359056806002144,\n",
 778         " 'ask': 0.0005359056806002144,\n",
 779         " 'audience': 0.0005359056806002144,\n",
 780         " 'begins': 0.0005359056806002144,\n",
 781         " 'behind': 0.0005359056806002144,\n",
 782         " 'believes': 0.0005359056806002144,\n",
 783         " 'bella': 0.0005359056806002144,\n",
 784         " 'bill': 0.0005359056806002144,\n",
 785         " 'blade': 0.0005359056806002144,\n",
 786         " 'book': 0.0005359056806002144,\n",
 787         " 'books': 0.0005359056806002144,\n",
 788         " 'brain': 0.0005359056806002144,\n",
 789         " 'bring': 0.0005359056806002144,\n",
 790         " 'camera': 0.0005359056806002144,\n",
 791         " 'care': 0.0005359056806002144,\n",
 792         " 'casey': 0.0005359056806002144,\n",
 793         " 'cell': 0.0005359056806002144,\n",
 794         " 'charlotte': 0.0005359056806002144,\n",
 795         " 'choice': 0.0005359056806002144,\n",
 796         " 'clear': 0.0005359056806002144,\n",
 797         " 'creature': 0.0005359056806002144,\n",
 798         " 'credit': 0.0005359056806002144,\n",
 799         " 'crew': 0.0005359056806002144,\n",
 800         " 'crowd': 0.0005359056806002144,\n",
 801         " 'days': 0.0005359056806002144,\n",
 802         " 'decides': 0.0005359056806002144,\n",
 803         " 'dennis': 0.0005359056806002144,\n",
 804         " 'die': 0.0005359056806002144,\n",
 805         " 'different': 0.0005359056806002144,\n",
 806         " 'doctor': 0.0005359056806002144,\n",
 807         " 'dream': 0.0005359056806002144,\n",
 808         " 'earth': 0.0005359056806002144,\n",
 809         " 'eat': 0.0005359056806002144,\n",
 810         " 'emma': 0.0005359056806002144,\n",
 811         " 'end': 0.0005359056806002144,\n",
 812         " 'ethan': 0.0005359056806002144,\n",
 813         " 'fear': 0.0005359056806002144,\n",
 814         " 'fire': 0.0005359056806002144,\n",
 815         " 'frankie': 0.0005359056806002144,\n",
 816         " 'gift': 0.0005359056806002144,\n",
 817         " 'great': 0.0005359056806002144,\n",
 818         " 'guards': 0.0005359056806002144,\n",
 819         " 'guns': 0.0005359056806002144,\n",
 820         " 'guys': 0.0005359056806002144,\n",
 821         " 'hair': 0.0005359056806002144,\n",
 822         " 'hands': 0.0005359056806002144,\n",
 823         " 'heads': 0.0005359056806002144,\n",
 824         " 'hell': 0.0005359056806002144,\n",
 825         " 'hurt': 0.0005359056806002144,\n",
 826         " 'immediately': 0.0005359056806002144,\n",
 827         " 'james': 0.0005359056806002144,\n",
 828         " 'jason': 0.0005359056806002144,\n",
 829         " 'jeff': 0.0005359056806002144,\n",
 830         " 'jenny': 0.0005359056806002144,\n",
 831         " 'jerry': 0.0005359056806002144,\n",
 832         " 'jimmy': 0.0005359056806002144,\n",
 833         " 'joy': 0.0005359056806002144,\n",
 834         " 'karen': 0.0005359056806002144,\n",
 835         " 'key': 0.0005359056806002144,\n",
 836         " 'keys': 0.0005359056806002144,\n",
 837         " 'kill': 0.0005359056806002144,\n",
 838         " 'killed': 0.0005359056806002144,\n",
 839         " 'king': 0.0005359056806002144,\n",
 840         " 'knows': 0.0005359056806002144,\n",
 841         " 'land': 0.0005359056806002144,\n",
 842         " 'lee': 0.0005359056806002144,\n",
 843         " 'leo': 0.0005359056806002144,\n",
 844         " 'let': 0.0005359056806002144,\n",
 845         " 'logan': 0.0005359056806002144,\n",
 846         " 'london': 0.0005359056806002144,\n",
 847         " 'lost': 0.0005359056806002144,\n",
 848         " 'machine': 0.0005359056806002144,\n",
 849         " 'major': 0.0005359056806002144,\n",
 850         " 'married': 0.0005359056806002144,\n",
 851         " 'mask': 0.0005359056806002144,\n",
 852         " 'means': 0.0005359056806002144,\n",
 853         " 'meanwhile': 0.0005359056806002144,\n",
 854         " 'meeting': 0.0005359056806002144,\n",
 855         " 'megan': 0.0005359056806002144,\n",
 856         " 'mexico': 0.0005359056806002144,\n",
 857         " 'mia': 0.0005359056806002144,\n",
 858         " 'miles': 0.0005359056806002144,\n",
 859         " 'mind': 0.0005359056806002144,\n",
 860         " 'miss': 0.0005359056806002144,\n",
 861         " 'missing': 0.0005359056806002144,\n",
 862         " 'mom': 0.0005359056806002144,\n",
 863         " 'must': 0.0005359056806002144,\n",
 864         " 'nate': 0.0005359056806002144,\n",
 865         " 'nearby': 0.0005359056806002144,\n",
 866         " 'ned': 0.0005359056806002144,\n",
 867         " 'new': 0.0005359056806002144,\n",
 868         " 'news': 0.0005359056806002144,\n",
 869         " 'note': 0.0005359056806002144,\n",
 870         " 'office': 0.0005359056806002144,\n",
 871         " 'opens': 0.0005359056806002144,\n",
 872         " 'pain': 0.0005359056806002144,\n",
 873         " 'past': 0.0005359056806002144,\n",
 874         " 'pictures': 0.0005359056806002144,\n",
 875         " 'quickly': 0.0005359056806002144,\n",
 876         " 'radio': 0.0005359056806002144,\n",
 877         " 'raj': 0.0005359056806002144,\n",
 878         " 'rich': 0.0005359056806002144,\n",
 879         " 'ring': 0.0005359056806002144,\n",
 880         " 'ron': 0.0005359056806002144,\n",
 881         " 'run': 0.0005359056806002144,\n",
 882         " 'running': 0.0005359056806002144,\n",
 883         " 'safe': 0.0005359056806002144,\n",
 884         " 'santa': 0.0005359056806002144,\n",
 885         " 'sean': 0.0005359056806002144,\n",
 886         " 'secret': 0.0005359056806002144,\n",
 887         " 'ship': 0.0005359056806002144,\n",
 888         " 'short': 0.0005359056806002144,\n",
 889         " 'sign': 0.0005359056806002144,\n",
 890         " 'simon': 0.0005359056806002144,\n",
 891         " 'smith': 0.0005359056806002144,\n",
 892         " 'sonny': 0.0005359056806002144,\n",
 893         " 'sound': 0.0005359056806002144,\n",
 894         " 'space': 0.0005359056806002144,\n",
 895         " 'stand': 0.0005359056806002144,\n",
 896         " 'stone': 0.0005359056806002144,\n",
 897         " 'suddenly': 0.0005359056806002144,\n",
 898         " 'susan': 0.0005359056806002144,\n",
 899         " 'teddy': 0.0005359056806002144,\n",
 900         " 'test': 0.0005359056806002144,\n",
 901         " 'tickets': 0.0005359056806002144,\n",
 902         " 'tina': 0.0005359056806002144,\n",
 903         " 'told': 0.0005359056806002144,\n",
 904         " 'tommy': 0.0005359056806002144,\n",
 905         " 'top': 0.0005359056806002144,\n",
 906         " 'training': 0.0005359056806002144,\n",
 907         " 'travis': 0.0005359056806002144,\n",
 908         " 'truck': 0.0005359056806002144,\n",
 909         " 'turn': 0.0005359056806002144,\n",
 910         " 'turns': 0.0005359056806002144,\n",
 911         " 'upset': 0.0005359056806002144,\n",
 912         " 'used': 0.0005359056806002144,\n",
 913         " 'van': 0.0005359056806002144,\n",
 914         " 'victor': 0.0005359056806002144,\n",
 915         " 'village': 0.0005359056806002144,\n",
 916         " 'waiting': 0.0005359056806002144,\n",
 917         " 'walk': 0.0005359056806002144,\n",
 918         " 'walter': 0.0005359056806002144,\n",
 919         " 'war': 0.0005359056806002144,\n",
 920         " 'went': 0.0005359056806002144,\n",
 921         " 'wish': 0.0005359056806002144,\n",
 922         " 'woody': 0.0005359056806002144,\n",
 923         " 'word': 0.0005359056806002144,\n",
 924         " 'words': 0.0005359056806002144,\n",
 925         " 'working': 0.0005359056806002144,\n",
 926         " 'yet': 0.0005359056806002144,\n",
 927         " 'young': 0.0005359056806002144,\n",
 928         " \"'d\": 0.0004976267034144848,\n",
 929         " 'actually': 0.0004976267034144848,\n",
 930         " 'admits': 0.0004976267034144848,\n",
 931         " 'agent': 0.0004976267034144848,\n",
 932         " 'alan': 0.0004976267034144848,\n",
 933         " 'almost': 0.0004976267034144848,\n",
 934         " 'angela': 0.0004976267034144848,\n",
 935         " 'angry': 0.0004976267034144848,\n",
 936         " 'animals': 0.0004976267034144848,\n",
 937         " 'anne': 0.0004976267034144848,\n",
 938         " 'anymore': 0.0004976267034144848,\n",
 939         " 'asking': 0.0004976267034144848,\n",
 940         " 'attack': 0.0004976267034144848,\n",
 941         " 'ball': 0.0004976267034144848,\n",
 942         " 'bank': 0.0004976267034144848,\n",
 943         " 'bar': 0.0004976267034144848,\n",
 944         " 'bed': 0.0004976267034144848,\n",
 945         " 'best': 0.0004976267034144848,\n",
 946         " 'bit': 0.0004976267034144848,\n",
 947         " 'bomb': 0.0004976267034144848,\n",
 948         " 'broken': 0.0004976267034144848,\n",
 949         " 'buddy': 0.0004976267034144848,\n",
 950         " 'calling': 0.0004976267034144848,\n",
 951         " 'camp': 0.0004976267034144848,\n",
 952         " 'candy': 0.0004976267034144848,\n",
 953         " 'captain': 0.0004976267034144848,\n",
 954         " 'career': 0.0004976267034144848,\n",
 955         " 'carol': 0.0004976267034144848,\n",
 956         " 'caroline': 0.0004976267034144848,\n",
 957         " 'catch': 0.0004976267034144848,\n",
 958         " 'charles': 0.0004976267034144848,\n",
 959         " 'clay': 0.0004976267034144848,\n",
 960         " 'contact': 0.0004976267034144848,\n",
 961         " 'continues': 0.0004976267034144848,\n",
 962         " 'conversation': 0.0004976267034144848,\n",
 963         " 'cooper': 0.0004976267034144848,\n",
 964         " 'country': 0.0004976267034144848,\n",
 965         " 'crash': 0.0004976267034144848,\n",
 966         " 'cross': 0.0004976267034144848,\n",
 967         " 'date': 0.0004976267034144848,\n",
 968         " 'doc': 0.0004976267034144848,\n",
 969         " 'doll': 0.0004976267034144848,\n",
 970         " 'done': 0.0004976267034144848,\n",
 971         " 'dress': 0.0004976267034144848,\n",
 972         " 'elizabeth': 0.0004976267034144848,\n",
 973         " 'evelyn': 0.0004976267034144848,\n",
 974         " 'evening': 0.0004976267034144848,\n",
 975         " 'face': 0.0004976267034144848,\n",
 976         " 'fall': 0.0004976267034144848,\n",
 977         " 'field': 0.0004976267034144848,\n",
 978         " 'floor': 0.0004976267034144848,\n",
 979         " 'force': 0.0004976267034144848,\n",
 980         " 'francis': 0.0004976267034144848,\n",
 981         " 'fred': 0.0004976267034144848,\n",
 982         " 'fresh': 0.0004976267034144848,\n",
 983         " 'fun': 0.0004976267034144848,\n",
 984         " 'girlfriend': 0.0004976267034144848,\n",
 985         " 'gives': 0.0004976267034144848,\n",
 986         " 'god': 0.0004976267034144848,\n",
 987         " 'government': 0.0004976267034144848,\n",
 988         " 'grace': 0.0004976267034144848,\n",
 989         " 'hank': 0.0004976267034144848,\n",
 990         " 'happy': 0.0004976267034144848,\n",
 991         " 'history': 0.0004976267034144848,\n",
 992         " 'holmes': 0.0004976267034144848,\n",
 993         " 'horse': 0.0004976267034144848,\n",
 994         " 'hour': 0.0004976267034144848,\n",
 995         " 'important': 0.0004976267034144848,\n",
 996         " 'insists': 0.0004976267034144848,\n",
 997         " 'interview': 0.0004976267034144848,\n",
 998         " 'j': 0.0004976267034144848,\n",
 999         " 'jamie': 0.0004976267034144848,\n",
1000         " 'jay': 0.0004976267034144848,\n",
1001         " 'jennifer': 0.0004976267034144848,\n",
1002         " 'jess': 0.0004976267034144848,\n",
1003         " 'joey': 0.0004976267034144848,\n",
1004         " 'jones': 0.0004976267034144848,\n",
1005         " 'jordan': 0.0004976267034144848,\n",
1006         " 'julia': 0.0004976267034144848,\n",
1007         " 'kathy': 0.0004976267034144848,\n",
1008         " 'kay': 0.0004976267034144848,\n",
1009         " 'kelly': 0.0004976267034144848,\n",
1010         " 'lara': 0.0004976267034144848,\n",
1011         " 'larry': 0.0004976267034144848,\n",
1012         " 'late': 0.0004976267034144848,\n",
1013         " 'lead': 0.0004976267034144848,\n",
1014         " 'learn': 0.0004976267034144848,\n",
1015         " 'leonard': 0.0004976267034144848,\n",
1016         " 'lets': 0.0004976267034144848,\n",
1017         " 'linda': 0.0004976267034144848,\n",
1018         " 'line': 0.0004976267034144848,\n",
1019         " 'listen': 0.0004976267034144848,\n",
1020         " 'lucas': 0.0004976267034144848,\n",
1021         " 'luke': 0.0004976267034144848,\n",
1022         " 'lunch': 0.0004976267034144848,\n",
1023         " 'made': 0.0004976267034144848,\n",
1024         " 'makes': 0.0004976267034144848,\n",
1025         " 'many': 0.0004976267034144848,\n",
1026         " 'map': 0.0004976267034144848,\n",
1027         " 'maria': 0.0004976267034144848,\n",
1028         " 'marriage': 0.0004976267034144848,\n",
1029         " 'martin': 0.0004976267034144848,\n",
1030         " 'may': 0.0004976267034144848,\n",
1031         " 'mickey': 0.0004976267034144848,\n",
1032         " 'minutes': 0.0004976267034144848,\n",
1033         " 'mirror': 0.0004976267034144848,\n",
1034         " 'mitch': 0.0004976267034144848,\n",
1035         " 'motel': 0.0004976267034144848,\n",
1036         " 'move': 0.0004976267034144848,\n",
1037         " 'music': 0.0004976267034144848,\n",
1038         " 'notice': 0.0004976267034144848,\n",
1039         " 'old': 0.0004976267034144848,\n",
1040         " 'orders': 0.0004976267034144848,\n",
1041         " 'penny': 0.0004976267034144848,\n",
1042         " 'pills': 0.0004976267034144848,\n",
1043         " 'planet': 0.0004976267034144848,\n",
1044         " 'play': 0.0004976267034144848,\n",
1045         " 'point': 0.0004976267034144848,\n",
1046         " 'present': 0.0004976267034144848,\n",
1047         " 'president': 0.0004976267034144848,\n",
1048         " 'priest': 0.0004976267034144848,\n",
1049         " 'prisoners': 0.0004976267034144848,\n",
1050         " 'private': 0.0004976267034144848,\n",
1051         " 'promises': 0.0004976267034144848,\n",
1052         " 'public': 0.0004976267034144848,\n",
1053         " 'put': 0.0004976267034144848,\n",
1054         " 'questions': 0.0004976267034144848,\n",
1055         " 'race': 0.0004976267034144848,\n",
1056         " 'rachel': 0.0004976267034144848,\n",
1057         " 'reason': 0.0004976267034144848,\n",
1058         " 'report': 0.0004976267034144848,\n",
1059         " 'restaurant': 0.0004976267034144848,\n",
1060         " 'revenge': 0.0004976267034144848,\n",
1061         " 'rock': 0.0004976267034144848,\n",
1062         " 'rope': 0.0004976267034144848,\n",
1063         " 'roy': 0.0004976267034144848,\n",
1064         " 'ryan': 0.0004976267034144848,\n",
1065         " 'safety': 0.0004976267034144848,\n",
1066         " 'scott': 0.0004976267034144848,\n",
1067         " 'search': 0.0004976267034144848,\n",
1068         " 'shane': 0.0004976267034144848,\n",
1069         " 'shows': 0.0004976267034144848,\n",
1070         " 'side': 0.0004976267034144848,\n",
1071         " 'situation': 0.0004976267034144848,\n",
1072         " 'sleep': 0.0004976267034144848,\n",
1073         " 'sleeping': 0.0004976267034144848,\n",
1074         " 'smile': 0.0004976267034144848,\n",
1075         " 'snow': 0.0004976267034144848,\n",
1076         " 'song': 0.0004976267034144848,\n",
1077         " 'sophie': 0.0004976267034144848,\n",
1078         " 'special': 0.0004976267034144848,\n",
1079         " 'spot': 0.0004976267034144848,\n",
1080         " 'staff': 0.0004976267034144848,\n",
1081         " 'stanley': 0.0004976267034144848,\n",
1082         " 'starts': 0.0004976267034144848,\n",
1083         " 'stolen': 0.0004976267034144848,\n",
1084         " 'stops': 0.0004976267034144848,\n",
1085         " 'sure': 0.0004976267034144848,\n",
1086         " 'surprise': 0.0004976267034144848,\n",
1087         " 'system': 0.0004976267034144848,\n",
1088         " 'tape': 0.0004976267034144848,\n",
1089         " 'taylor': 0.0004976267034144848,\n",
1090         " 'thing': 0.0004976267034144848,\n",
1091         " 'think': 0.0004976267034144848,\n",
1092         " 'toby': 0.0004976267034144848,\n",
1093         " 'together': 0.0004976267034144848,\n",
1094         " 'tracy': 0.0004976267034144848,\n",
1095         " 'train': 0.0004976267034144848,\n",
1096         " 'trio': 0.0004976267034144848,\n",
1097         " 'tv': 0.0004976267034144848,\n",
1098         " 'uncle': 0.0004976267034144848,\n",
1099         " 'using': 0.0004976267034144848,\n",
1100         " 'vincent': 0.0004976267034144848,\n",
1101         " 'walking': 0.0004976267034144848,\n",
1102         " 'walks': 0.0004976267034144848,\n",
1103         " 'wall': 0.0004976267034144848,\n",
1104         " 'weekend': 0.0004976267034144848,\n",
1105         " 'worse': 0.0004976267034144848,\n",
1106         " 'younger': 0.0004976267034144848,\n",
1107         " '2': 0.00045934772622875517,\n",
1108         " 'ace': 0.00045934772622875517,\n",
1109         " 'action': 0.00045934772622875517,\n",
1110         " 'address': 0.00045934772622875517,\n",
1111         " 'al': 0.00045934772622875517,\n",
1112         " 'alarm': 0.00045934772622875517,\n",
1113         " 'albert': 0.00045934772622875517,\n",
1114         " 'alfred': 0.00045934772622875517,\n",
1115         " 'america': 0.00045934772622875517,\n",
1116         " 'angel': 0.00045934772622875517,\n",
1117         " 'anyway': 0.00045934772622875517,\n",
1118         " 'arm': 0.00045934772622875517,\n",
1119         " 'arrive': 0.00045934772622875517,\n",
1120         " 'arrives': 0.00045934772622875517,\n",
1121         " 'assistant': 0.00045934772622875517,\n",
1122         " 'barbara': 0.00045934772622875517,\n",
1123         " 'barry': 0.00045934772622875517,\n",
1124         " 'base': 0.00045934772622875517,\n",
1125         " 'beach': 0.00045934772622875517,\n",
1126         " 'bedroom': 0.00045934772622875517,\n",
1127         " 'beer': 0.00045934772622875517,\n",
1128         " 'believe': 0.00045934772622875517,\n",
1129         " 'bet': 0.00045934772622875517,\n",
1130         " 'bike': 0.00045934772622875517,\n",
1131         " 'birthday': 0.00045934772622875517,\n",
1132         " 'boss': 0.00045934772622875517,\n",
1133         " 'bottle': 0.00045934772622875517,\n",
1134         " 'buzz': 0.00045934772622875517,\n",
1135         " 'campbell': 0.00045934772622875517,\n",
1136         " 'cards': 0.00045934772622875517,\n",
1137         " 'carla': 0.00045934772622875517,\n",
1138         " 'carrie': 0.00045934772622875517,\n",
1139         " 'cars': 0.00045934772622875517,\n",
1140         " 'chair': 0.00045934772622875517,\n",
1141         " 'character': 0.00045934772622875517,\n",
1142         " 'chicago': 0.00045934772622875517,\n",
1143         " 'chief': 0.00045934772622875517,\n",
1144         " 'christine': 0.00045934772622875517,\n",
1145         " 'cindy': 0.00045934772622875517,\n",
1146         " 'club': 0.00045934772622875517,\n",
1147         " 'cocaine': 0.00045934772622875517,\n",
1148         " 'comes': 0.00045934772622875517,\n",
1149         " 'coming': 0.00045934772622875517,\n",
1150         " 'computer': 0.00045934772622875517,\n",
1151         " 'courage': 0.00045934772622875517,\n",
1152         " 'court': 0.00045934772622875517,\n",
1153         " 'customers': 0.00045934772622875517,\n",
1154         " 'dale': 0.00045934772622875517,\n",
1155         " 'damage': 0.00045934772622875517,\n",
1156         " 'dan': 0.00045934772622875517,\n",
1157         " 'dangerous': 0.00045934772622875517,\n",
1158         " 'dark': 0.00045934772622875517,\n",
1159         " 'debbie': 0.00045934772622875517,\n",
1160         " 'demands': 0.00045934772622875517,\n",
1161         " 'details': 0.00045934772622875517,\n",
1162         " 'diamonds': 0.00045934772622875517,\n",
1163         " 'dies': 0.00045934772622875517,\n",
1164         " 'distance': 0.00045934772622875517,\n",
1165         " 'documents': 0.00045934772622875517,\n",
1166         " 'dogs': 0.00045934772622875517,\n",
1167         " 'dorothy': 0.00045934772622875517,\n",
1168         " 'doug': 0.00045934772622875517,\n",
1169         " 'dreams': 0.00045934772622875517,\n",
1170         " 'driving': 0.00045934772622875517,\n",
1171         " 'drops': 0.00045934772622875517,\n",
1172         " 'drug': 0.00045934772622875517,\n",
1173         " 'drunk': 0.00045934772622875517,\n",
1174         " 'due': 0.00045934772622875517,\n",
1175         " 'duke': 0.00045934772622875517,\n",
1176         " 'edward': 0.00045934772622875517,\n",
1177         " 'elena': 0.00045934772622875517,\n",
1178         " 'eric': 0.00045934772622875517,\n",
1179         " 'eve': 0.00045934772622875517,\n",
1180         " 'everybody': 0.00045934772622875517,\n",
1181         " 'experience': 0.00045934772622875517,\n",
1182         " 'eye': 0.00045934772622875517,\n",
1183         " 'eyes': 0.00045934772622875517,\n",
1184         " 'fake': 0.00045934772622875517,\n",
1185         " 'falls': 0.00045934772622875517,\n",
1186         " 'fast': 0.00045934772622875517,\n",
1187         " 'fights': 0.00045934772622875517,\n",
1188         " 'five': 0.00045934772622875517,\n",
1189         " 'following': 0.00045934772622875517,\n",
1190         " 'follows': 0.00045934772622875517,\n",
1191         " 'foot': 0.00045934772622875517,\n",
1192         " 'found': 0.00045934772622875517,\n",
1193         " 'four': 0.00045934772622875517,\n",
1194         " 'fox': 0.00045934772622875517,\n",
1195         " 'front': 0.00045934772622875517,\n",
1196         " 'furious': 0.00045934772622875517,\n",
1197         " 'future': 0.00045934772622875517,\n",
1198         " 'garage': 0.00045934772622875517,\n",
1199         " 'general': 0.00045934772622875517,\n",
1200         " 'ghost': 0.00045934772622875517,\n",
1201         " 'grant': 0.00045934772622875517,\n",
1202         " 'gus': 0.00045934772622875517,\n",
1203         " 'half': 0.00045934772622875517,\n",
1204         " 'hall': 0.00045934772622875517,\n",
1205         " 'harold': 0.00045934772622875517,\n",
1206         " 'heard': 0.00045934772622875517,\n",
1207         " 'hit': 0.00045934772622875517,\n",
1208         " 'hold': 0.00045934772622875517,\n",
1209         " 'howard': 0.00045934772622875517,\n",
1210         " 'human': 0.00045934772622875517,\n",
1211         " 'identity': 0.00045934772622875517,\n",
1212         " 'ill': 0.00045934772622875517,\n",
1213         " 'island': 0.00045934772622875517,\n",
1214         " 'ivan': 0.00045934772622875517,\n",
1215         " 'jail': 0.00045934772622875517,\n",
1216         " 'joke': 0.00045934772622875517,\n",
1217         " 'jonathan': 0.00045934772622875517,\n",
1218         " 'josh': 0.00045934772622875517,\n",
1219         " 'journey': 0.00045934772622875517,\n",
1220         " 'jules': 0.00045934772622875517,\n",
1221         " 'jumps': 0.00045934772622875517,\n",
1222         " 'katherine': 0.00045934772622875517,\n",
1223         " 'kim': 0.00045934772622875517,\n",
1224         " 'kirk': 0.00045934772622875517,\n",
1225         " 'kiss': 0.00045934772622875517,\n",
1226         " 'kit': 0.00045934772622875517,\n",
1227         " 'kitchen': 0.00045934772622875517,\n",
1228         " 'law': 0.00045934772622875517,\n",
1229         " 'less': 0.00045934772622875517,\n",
1230         " 'lightning': 0.00045934772622875517,\n",
1231         " 'likely': 0.00045934772622875517,\n",
1232         " 'lily': 0.00045934772622875517,\n",
1233         " 'list': 0.00045934772622875517,\n",
1234         " 'living': 0.00045934772622875517,\n",
1235         " 'lizzie': 0.00045934772622875517,\n",
1236         " 'local': 0.00045934772622875517,\n",
1237         " 'longer': 0.00045934772622875517,\n",
1238         " 'lori': 0.00045934772622875517,\n",
1239         " 'making': 0.00045934772622875517,\n",
1240         " 'malcolm': 0.00045934772622875517,\n",
1241         " 'marcus': 0.00045934772622875517,\n",
1242         " 'martha': 0.00045934772622875517,\n",
1243         " 'mason': 0.00045934772622875517,\n",
1244         " 'match': 0.00045934772622875517,\n",
1245         " 'matter': 0.00045934772622875517,\n",
1246         " 'maybe': 0.00045934772622875517,\n",
1247         " 'medicine': 0.00045934772622875517,\n",
1248         " 'mission': 0.00045934772622875517,\n",
1249         " 'molly': 0.00045934772622875517,\n",
1250         " 'moses': 0.00045934772622875517,\n",
1251         " 'moves': 0.00045934772622875517,\n",
1252         " 'murder': 0.00045934772622875517,\n",
1253         " 'nancy': 0.00045934772622875517,\n",
1254         " 'nathan': 0.00045934772622875517,\n",
1255         " 'nose': 0.00045934772622875517,\n",
1256         " 'offering': 0.00045934772622875517,\n",
1257         " 'offers': 0.00045934772622875517,\n",
1258         " 'officers': 0.00045934772622875517,\n",
1259         " 'oliver': 0.00045934772622875517,\n",
1260         " 'open': 0.00045934772622875517,\n",
1261         " 'paid': 0.00045934772622875517,\n",
1262         " 'painting': 0.00045934772622875517,\n",
1263         " 'partner': 0.00045934772622875517,\n",
1264         " 'patients': 0.00045934772622875517,\n",
1265         " 'peace': 0.00045934772622875517,\n",
1266         " 'performance': 0.00045934772622875517,\n",
1267         " 'photos': 0.00045934772622875517,\n",
1268         " 'piece': 0.00045934772622875517,\n",
1269         " 'pilot': 0.00045934772622875517,\n",
1270         " 'pistol': 0.00045934772622875517,\n",
1271         " 'plane': 0.00045934772622875517,\n",
1272         " 'press': 0.00045934772622875517,\n",
1273         " 'price': 0.00045934772622875517,\n",
1274         " 'product': 0.00045934772622875517,\n",
1275         " 'protect': 0.00045934772622875517,\n",
1276         " 'pulls': 0.00045934772622875517,\n",
1277         " 'queen': 0.00045934772622875517,\n",
1278         " 'realize': 0.00045934772622875517,\n",
1279         " 'rebecca': 0.00045934772622875517,\n",
1280         " 'record': 0.00045934772622875517,\n",
1281         " 'records': 0.00045934772622875517,\n",
1282         " 'remember': 0.00045934772622875517,\n",
1283         " 'request': 0.00045934772622875517,\n",
1284         " 'respect': 0.00045934772622875517,\n",
1285         " 'richard': 0.00045934772622875517,\n",
1286         " 'ride': 0.00045934772622875517,\n",
1287         " 'road': 0.00045934772622875517,\n",
1288         " 'robin': 0.00045934772622875517,\n",
1289         " 'rocky': 0.00045934772622875517,\n",
1290         " 'role': 0.00045934772622875517,\n",
1291         " 'roof': 0.00045934772622875517,\n",
1292         " 'ruth': 0.00045934772622875517,\n",
1293         " 'said': 0.00045934772622875517,\n",
1294         " 'samantha': 0.00045934772622875517,\n",
1295         " 'sandy': 0.00045934772622875517,\n",
1296         " 'seems': 0.00045934772622875517,\n",
1297         " 'serious': 0.00045934772622875517,\n",
1298         " 'shelter': 0.00045934772622875517,\n",
1299         " 'shit': 0.00045934772622875517,\n",
1300         " 'shoes': 0.00045934772622875517,\n",
1301         " 'shoot': 0.00045934772622875517,\n",
1302         " 'shooting': 0.00045934772622875517,\n",
1303         " 'shotgun': 0.00045934772622875517,\n",
1304         " 'shut': 0.00045934772622875517,\n",
1305         " 'small': 0.00045934772622875517,\n",
1306         " 'snake': 0.00045934772622875517,\n",
1307         " 'spare': 0.00045934772622875517,\n",
1308         " 'speech': 0.00045934772622875517,\n",
1309         " 'spirit': 0.00045934772622875517,\n",
1310         " 'stage': 0.00045934772622875517,\n",
1311         " 'star': 0.00045934772622875517,\n",
1312         " 'state': 0.00045934772622875517,\n",
1313         " 'storm': 0.00045934772622875517,\n",
1314         " 'strange': 0.00045934772622875517,\n",
1315         " 'students': 0.00045934772622875517,\n",
1316         " 'stuff': 0.00045934772622875517,\n",
1317         " 'suggestion': 0.00045934772622875517,\n",
1318         " 'suit': 0.00045934772622875517,\n",
1319         " 'supplies': 0.00045934772622875517,\n",
1320         " 'support': 0.00045934772622875517,\n",
1321         " 'taken': 0.00045934772622875517,\n",
1322         " 'ted': 0.00045934772622875517,\n",
1323         " 'telling': 0.00045934772622875517,\n",
1324         " 'thomas': 0.00045934772622875517,\n",
1325         " 'ticket': 0.00045934772622875517,\n",
1326         " 'tim': 0.00045934772622875517,\n",
1327         " 'times': 0.00045934772622875517,\n",
1328         " 'todd': 0.00045934772622875517,\n",
1329         " 'tree': 0.00045934772622875517,\n",
1330         " 'trip': 0.00045934772622875517,\n",
1331         " 'troy': 0.00045934772622875517,\n",
1332         " 'truth': 0.00045934772622875517,\n",
1333         " 'vampire': 0.00045934772622875517,\n",
1334         " 'visit': 0.00045934772622875517,\n",
1335         " 'wake': 0.00045934772622875517,\n",
1336         " 'watson': 0.00045934772622875517,\n",
1337         " 'ways': 0.00045934772622875517,\n",
1338         " 'week': 0.00045934772622875517,\n",
1339         " 'weeks': 0.00045934772622875517,\n",
1340         " 'william': 0.00045934772622875517,\n",
1341         " 'wilson': 0.00045934772622875517,\n",
1342         " 'wolf': 0.00045934772622875517,\n",
1343         " 'woods': 0.00045934772622875517,\n",
1344         " '1': 0.00042106874904302555,\n",
1345         " 'aaron': 0.00042106874904302555,\n",
1346         " 'abby': 0.00042106874904302555,\n",
1347         " 'advantage': 0.00042106874904302555,\n",
1348         " 'afraid': 0.00042106874904302555,\n",
1349         " 'agree': 0.00042106874904302555,\n",
1350         " 'alcohol': 0.00042106874904302555,\n",
1351         " 'allison': 0.00042106874904302555,\n",
1352         " 'although': 0.00042106874904302555,\n",
1353         " 'always': 0.00042106874904302555,\n",
1354         " 'ambulance': 0.00042106874904302555,\n",
1355         " 'ann': 0.00042106874904302555,\n",
1356         " 'anthony': 0.00042106874904302555,\n",
1357         " 'archer': 0.00042106874904302555,\n",
1358         " 'arms': 0.00042106874904302555,\n",
1359         " 'around': 0.00042106874904302555,\n",
1360         " 'arrested': 0.00042106874904302555,\n",
1361         " 'assignment': 0.00042106874904302555,\n",
1362         " 'attacked': 0.00042106874904302555,\n",
1363         " 'attempts': 0.00042106874904302555,\n",
1364         " 'band': 0.00042106874904302555,\n",
1365         " 'basement': 0.00042106874904302555,\n",
1366         " 'bath': 0.00042106874904302555,\n",
1367         " 'bathroom': 0.00042106874904302555,\n",
1368         " 'batman': 0.00042106874904302555,\n",
1369         " 'bear': 0.00042106874904302555,\n",
1370         " 'begin': 0.00042106874904302555,\n",
1371         " 'bishop': 0.00042106874904302555,\n",
1372         " 'bodies': 0.00042106874904302555,\n",
1373         " 'booth': 0.00042106874904302555,\n",
1374         " 'boyfriend': 0.00042106874904302555,\n",
1375         " 'bracelet': 0.00042106874904302555,\n",
1376         " 'brad': 0.00042106874904302555,\n",
1377         " 'breakfast': 0.00042106874904302555,\n",
1378         " 'brenda': 0.00042106874904302555,\n",
1379         " 'bullet': 0.00042106874904302555,\n",
1380         " 'buy': 0.00042106874904302555,\n",
1381         " 'cab': 0.00042106874904302555,\n",
1382         " 'cake': 0.00042106874904302555,\n",
1383         " 'carter': 0.00042106874904302555,\n",
1384         " 'castle': 0.00042106874904302555,\n",
1385         " 'charge': 0.00042106874904302555,\n",
1386         " 'charges': 0.00042106874904302555,\n",
1387         " 'chest': 0.00042106874904302555,\n",
1388         " 'chip': 0.00042106874904302555,\n",
1389         " 'chloe': 0.00042106874904302555,\n",
1390         " 'christian': 0.00042106874904302555,\n",
1391         " 'christopher': 0.00042106874904302555,\n",
1392         " 'claim': 0.00042106874904302555,\n",
1393         " 'claims': 0.00042106874904302555,\n",
1394         " 'classes': 0.00042106874904302555,\n",
1395         " 'clean': 0.00042106874904302555,\n",
1396         " 'cliff': 0.00042106874904302555,\n",
1397         " 'close': 0.00042106874904302555,\n",
1398         " 'coat': 0.00042106874904302555,\n",
1399         " 'coffee': 0.00042106874904302555,\n",
1400         " 'coin': 0.00042106874904302555,\n",
1401         " 'college': 0.00042106874904302555,\n",
1402         " 'complete': 0.00042106874904302555,\n",
1403         " 'continue': 0.00042106874904302555,\n",
1404         " 'convince': 0.00042106874904302555,\n",
1405         " 'cop': 0.00042106874904302555,\n",
1406         " 'danger': 0.00042106874904302555,\n",
1407         " 'darkness': 0.00042106874904302555,\n",
1408         " 'decided': 0.00042106874904302555,\n",
1409         " 'decision': 0.00042106874904302555,\n",
1410         " 'derek': 0.00042106874904302555,\n",
1411         " 'detective': 0.00042106874904302555,\n",
1412         " 'discovers': 0.00042106874904302555,\n",
1413         " 'donna': 0.00042106874904302555,\n",
1414         " 'donnie': 0.00042106874904302555,\n",
1415         " 'downstairs': 0.00042106874904302555,\n",
1416         " 'dracula': 0.00042106874904302555,\n",
1417         " ...}"
1418        ]
1419       },
1420       "execution_count": 28,
1421       "metadata": {},
1422       "output_type": "execute_result"
1423      }
1424     ],
1425     "source": [
1426      "find_most_probable('money')"
1427     ]
1428    },
1429    {
1430     "cell_type": "code",
1431     "execution_count": null,
1432     "id": "898438a3",
1433     "metadata": {},
1434     "outputs": [],
1435     "source": [
1436      "14/20"
1437     ]
1438    },
1439    {
1440     "cell_type": "code",
1441     "execution_count": null,
1442     "id": "b82f46fe",
1443     "metadata": {},
1444     "outputs": [],
1445     "source": [
1446      "a=count_f(words)"
1447     ]
1448    },
1449    {
1450     "cell_type": "code",
1451     "execution_count": null,
1452     "id": "ce8ff4e2",
1453     "metadata": {},
1454     "outputs": [],
1455     "source": []
1456    }
1457   ],
1458   "metadata": {
1459    "kernelspec": {
1460     "display_name": "Python 3",
1461     "language": "python",
1462     "name": "python3"
1463    },
1464    "language_info": {
1465     "codemirror_mode": {
1466      "name": "ipython",
1467      "version": 3
1468     },
1469     "file_extension": ".py",
1470     "mimetype": "text/x-python",
1471     "name": "python",
1472     "nbconvert_exporter": "python",
1473     "pygments_lexer": "ipython3",
1474     "version": "3.8.8"
1475    }
1476   },
1477   "nbformat": 4,
1478   "nbformat_minor": 5
1479  }
1480