{ "metadata": { "name": "", "signature": "sha256:ca858311da3d1741c2abd43d545452b7d0d5674d31288dbbb4a908f66e193e36" }, "nbformat": 3, "nbformat_minor": 0, "worksheets": [ { "cells": [ { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "Load Gadsby and create dataframe of frequency distribution" ] }, { "cell_type": "code", "collapsed": false, "input": [ "import re\n", "import os\n", "import nltk\n", "raw = open('gadsby_full_lower.txt', 'r').read()\n", "tokens = nltk.word_tokenize(raw)\n", "text = nltk.Text(tokens)\n", "\n", "# summary statistics\n", "print ' '\n", "print \"Characters: {}\".format(len(raw))\n", "print \"Tokens: {}\".format(len(tokens))\n", "print \"Unique tokens: {}\".format(len(set(tokens)))\n", "print \"Lexical diversity: {:.3f}\".format(len(set(tokens))*1.0/len(tokens))\n", "\n", "# create frequency distribution dataframe\n", "fdist = nltk.FreqDist(text)\n", "import pandas as pd\n", "df = pd.DataFrame({'token': fdist.keys(), 'freq':fdist.values()})\n", "df = df[~df.token.str.contains('[^A-Za-z]')] # remove tokens with non-alphabetic characters\n", "df.columns = ['freq', 'word']\n", "print ' '\n", "print df.head()" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ " \n", "Characters: 277418\n", "Tokens: 61759\n", "Unique tokens: 5233\n", "Lexical diversity: 0.085\n", " " ] }, { "output_type": "stream", "stream": "stdout", "text": [ "\n", " freq word\n", "1 2437 a\n", "2 1669 and\n", "4 1225 that\n", "6 1162 of\n", "8 934 in\n" ] } ], "prompt_number": 1 }, { "cell_type": "code", "collapsed": false, "input": [ "#sanity check\n", "print df[df.word.str.contains('e')]" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "Empty DataFrame\n", "Columns: [freq, word]\n", "Index: []\n" ] } ], "prompt_number": 2 }, { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "Make frequency dataframe of Brown corpus words not containing 'e'" ] }, { "cell_type": "code", "collapsed": false, "input": [ "if not os.path.isfile('brown_df.pickle'):\n", " print \"Processing Brown corpus from NLTK\"\n", " from nltk.corpus import brown\n", " categories = []\n", " words = []\n", " frequencies = []\n", " for category in brown.categories():\n", " wordlist = brown.words(categories=category)\n", " freqs = nltk.FreqDist([w.lower() for w in wordlist])\n", " for key in freqs.keys():\n", " categories.append(category)\n", " words.append(key)\n", " frequencies.append(freqs[key])\n", " brown_df = pd.DataFrame({'word':words, 'freq':frequencies, 'category':categories})\n", " brown_df['nonalpha'] = False\n", " brown_df['nonalpha'][brown_df.word.str.contains('[^A-Za-z]')] = True\n", " brown_df.to_pickle('brown_df.pickle')\n", "else:\n", " print \"Reading brown_df.pickle\"\n", " brown_df = pd.read_pickle('brown_df.pickle')" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "Reading brown_df.pickle\n" ] } ], "prompt_number": 3 }, { "cell_type": "code", "collapsed": false, "input": [ "if not os.path.isfile('brown_non_e.pickle'):\n", " print 'Creating dataframe.'\n", " total_freq = 0\n", " weighted_length = 0\n", " brown_words = brown_df[brown_df.nonalpha == False]\n", " brown_words = brown_words[~brown_words.word.str.contains('e')]\n", " brown_words = pd.DataFrame(brown_words.groupby(['word']).sum()).reset_index(drop=False)\n", " brown_words['length'] = 0\n", " brown_words.sort('freq', ascending = False, inplace=True)\n", " total_freq = brown_words.freq.sum()\n", " median_counter = total_freq / 2\n", " median_found = False\n", " for idx, row in brown_words.iterrows():\n", " curr_len = len(brown_words.word[idx])\n", " brown_words.length[idx] = curr_len\n", " weighted_length += brown_words.freq[idx] * curr_len\n", " median_counter -= curr_len * brown_words.freq[idx]\n", " if median_counter < 0 and median_found == False:\n", " wt_median_len = curr_len\n", " median_found = True\n", " brown_words = brown_words[['word', 'freq', 'length']]\n", " brown_words.to_pickle('brown_non_e.pickle')\n", " \n", "else:\n", " print 'Reading pickle.'\n", " brown_words = pd.read_pickle('brown_non_e.pickle')" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "Reading pickle.\n" ] } ], "prompt_number": 4 }, { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "Determine log likelihood of word frequencies, comparing Gadsby to Brown-without-e" ] }, { "cell_type": "code", "collapsed": false, "input": [ "def loglike(n1, t1, n2, t2):\n", " \"\"\"Calculates Dunning log likelihood of an observation of \n", " frequency n1 in a corpus of size t1, compared to a frequency n2 \n", " in a corpus of size t2. If result is positive, it is more \n", " likely to occur in corpus 1, otherwise in corpus 2.\"\"\"\n", " e1 = t1*1.0*(n1+n2)/(t1+t2) # expected values\n", " e2 = t2*1.0*(n1+n2)/(t1+t2)\n", " LL = 2 * ((n1 * log(n1/e1)) + n2 * (log(n2/e2)))\n", " if n2*1.0/t2 > n1*1.0/t1:\n", " LL = -LL\n", " return LL" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 5 }, { "cell_type": "code", "collapsed": false, "input": [ "from numpy import log\n", "\n", "t1 = df.freq.sum()\n", "t2 = brown_words.freq.sum()\n", "df['log_likelihood'] = 0.0\n", "\n", "for i in range(len(df)):\n", " word = df.word.iloc[i]\n", " n1 = df.freq.iloc[i]\n", " fnd = brown_words[brown_words.word == word]\n", " if len(fnd) > 0:\n", " n2 = fnd.freq.iloc[0]\n", " else:\n", " n2 = 0\n", " df.log_likelihood.iloc[i] = loglike(n1, t1, n2, t2)" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 6 }, { "cell_type": "code", "collapsed": false, "input": [ "print df.sort('log_likelihood', ascending=False).head(20)" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ " freq word log_likelihood\n", "54 200 hills 749.660964\n", "32 297 big 630.151174\n", "79 122 nancy 563.296979\n", "71 138 lady 411.288896\n", "68 140 bill 326.549776\n", "108 80 folks 306.118227\n", "95 97 honor 272.219017\n", "59 187 young 271.193850\n", "85 108 happy 267.283070\n", "112 76 sarah 265.372978\n", "44 237 old 253.839785\n", "136 60 simpkins 253.691283\n", "19 442 so 234.694974\n", "47 229 know 226.664917\n", "152 52 councilman 224.677398\n", "176 46 nina 218.448458\n", "87 106 girls 211.621005\n", "126 66 tiny 177.231911\n", "103 89 oh 175.102218\n", "131 63 grand 168.765580\n" ] } ], "prompt_number": 7 }, { "cell_type": "code", "collapsed": false, "input": [ "print df.sort('log_likelihood', ascending=True).head(20)" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ " freq word log_likelihood\n", "6 1162 of -1776.113782\n", "9 921 to -1112.090217\n", "8 934 in -616.702852\n", "2 1669 and -383.521998\n", "78 124 by -362.344996\n", "18 471 is -254.363544\n", "22 403 with -113.644883\n", "25 383 his -112.694084\n", "96 96 him -104.650941\n", "104 86 has -103.254890\n", "13 585 for -97.381118\n", "35 284 at -96.568517\n", "24 384 on -94.773170\n", "228 34 may -92.570425\n", "30 301 had -64.783953\n", "4378 1 program -60.080679\n", "2736 1 among -55.998265\n", "247 30 must -54.226455\n", "41 259 from -52.902273\n", "33 291 not -42.458054\n" ] } ], "prompt_number": 8 }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Complete analysis dataframe\n", "\n", "* Add column of Brown non-e frequency normalized to Gadsby corpus length\n", "* Add absolute and relative frequency differences between Gasby and Brown non-e\n", "* Add frequency of neighboring 'the' or 'e'-containing word in original Brown corpus\n", "\n", "*Note: this takes a long time.*" ] }, { "cell_type": "code", "collapsed": false, "input": [ "import time\n", "import numpy as np\n", "from nltk.corpus import brown\n", "br_trig = nltk.trigrams(brown.words())\n", "\n", "if 2>1: #not os.path.isfile('gadsby_analysis.pickle'):\n", " start = time.time()\n", " print \"Building dataframe.\"\n", "\n", " freq_g = df.freq.sum()\n", " freq_b = brown_words.freq.sum()\n", "\n", " df['brown_freq_normalized'] = 0.0\n", " for i in range(len(df)):\n", " word = df.word.iloc[i]\n", " try:\n", " brown_freq = brown_words[brown_words.word == word].freq.iloc[0]\n", " except:\n", " brown_freq = 0\n", " try:\n", " df.brown_freq_normalized.iloc[i] = brown_freq * freq_g / freq_b\n", " except:\n", " df.brown_freq_normalized.iloc[i] = np.nan\n", "\n", " df['diff_absolute'] = df.freq - df.brown_freq_normalized\n", " df['diff_relative'] = df.freq*1.0/df.brown_freq_normalized\n", " \n", " wprev = []\n", " word = []\n", " wnext = []\n", " for item in br_trig:\n", " a = item[0].lower()\n", " b = item[1].lower()\n", " c = item[2].lower()\n", " if (re.search('[a-z]', a) and\n", " re.search('[a-z]', b) and\n", " re.search('[a-z]', c) ):\n", " wprev.append(a)\n", " word.append(b)\n", " wnext.append(c)\n", "\n", " tri = pd.DataFrame({'wprev': wprev,\n", " 'word': word,\n", " 'wnext': wnext})\n", "\n", " def calc_pcte(row):\n", " dftemp = tri[tri.word == row.word]\n", " total = len(dftemp)\n", " dftempe = dftemp[(dftemp.wprev.str.contains('e') ) | (dftemp.wnext.str.contains('e'))]\n", " try:\n", " return len(dftempe) * 100.0 / total\n", " except:\n", " return np.nan\n", "\n", " def calc_pctthe(row):\n", " dftemp = tri[tri.word == row.word]\n", " total = len(dftemp)\n", " dftempthe = dftemp[(dftemp.wprev == 'the' ) | (dftemp.wnext == 'the')]\n", " try:\n", " return len(dftempthe) * 100.0 / total\n", " except:\n", " return np.nan\n", "\n", " df['pct_e'] = df.apply(calc_pcte, axis=1)\n", " df['pct_the'] = df.apply(calc_pctthe, axis=1)\n", "\n", " df.to_pickle('gadsby_analysis.pickle')\n", " df.to_csv('gadsby_analysis.csv')\n", " print \"Done. {} minutes elapsed.\".format(round((time.time() - start) / 60, 1))\n", " \n", "else:\n", " print \"Reading pickle.\"\n", " df = pd.read_pickle('gadsby_analysis.pickle')" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "Building dataframe.\n", "Done. 17.2 minutes elapsed." ] }, { "output_type": "stream", "stream": "stdout", "text": [ "\n" ] } ], "prompt_number": 13 }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "\n", "Graph 'volcano plot' of " ] }, { "cell_type": "code", "collapsed": false, "input": [ "from matplotlib import pyplot as plt\n", "%matplotlib inline\n", "import numpy as np\n", "import seaborn\n", "\n", "plt.figure(figsize=(9,6))\n", "\n", "dftemp = df[['word', 'log_likelihood', 'pct_e', 'freq']].dropna()\n", "#dftemp = dftemp[(dftemp.log_likelihood > -50) & (dftemp.log_likelihood < 50)]\n", "dftemp = dftemp[(dftemp.freq > 10)]\n", "\n", "plt.scatter(dftemp.log_likelihood, dftemp.pct_e, s=dftemp.freq*1.5, marker='o', \n", " color='blue', alpha=0.15, label='data')\n", "plt.title(' ')\n", "#plt.xlim(-10,10)\n", "plt.xlabel('Log-Likelihood that word is overrepresented in Gadsby')\n", "plt.ylabel(\"Percent probability of neighbouring 'e' in Brown corpus\")\n", "\n", "plt.show()\n" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAjUAAAGACAYAAABV+pPDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n", "text": [ "" ] } ], "prompt_number": 14 }, { "cell_type": "code", "collapsed": false, "input": [ "df" ], "language": "python", "metadata": {}, "outputs": [ { "html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
freqwordlog_likelihoodbrown_freq_normalizeddiff_absolutediff_relative
1 2437 a 39.837277 2125 312 1.146824
2 1669 and -383.521998 2643 -974 0.631479
4 1225 that 55.933699 970 255 1.262887
6 1162 of-1776.113782 3336-2174 0.348321
8 934 in -616.702852 1954-1020 0.477994
9 921 to-1112.090217 2396-1475 0.384391
10 710 was -39.608135 899 -189 0.789766
11 694 as 1.177540 664 30 1.045181
12 667 it -22.394679 802 -135 0.831671
13 585 for -97.381118 869 -284 0.673188
14 566 you 165.609370 301 265 1.880399
15 518 but 28.177335 401 117 1.291771
17 471 i -0.008889 473 -2 0.995772
18 471 is -254.363544 926 -455 0.508639
19 442 so 234.694974 181 261 2.441989
21 421 this -5.137938 471 -50 0.893843
22 403 with -113.644883 667 -264 0.604198
23 393 all 40.463187 274 119 1.434307
24 384 on -94.773170 617 -233 0.622366
25 383 his -112.694084 640 -257 0.598437
26 364 gadsby NaN 0 364 inf
28 327 up 96.888501 173 154 1.890173
30 301 had -64.783953 470 -169 0.640426
31 298 or -19.827346 385 -87 0.774026
32 297 big 630.151174 32 265 9.281250
33 291 not -42.458054 422 -131 0.689573
34 289 an -8.174090 342 -53 0.845029
35 284 at -96.568517 492 -208 0.577236
36 280 out 31.832340 192 88 1.458333
37 270 our 134.651348 114 156 2.368421
.....................
5122 1 wistaria NaN 0 1 inf
5124 1 wit -0.421448 1 0 1.000000
5126 1 withhold 1.487055 0 1 inf
5128 1 wobbly 1.487055 0 1 inf
5129 1 woild NaN 0 1 inf
5130 1 wolf 0.265836 0 1 inf
5133 1 wondrous 2.358220 0 1 inf
5134 1 woo 0.982790 0 1 inf
5137 1 woodlands NaN 0 1 inf
5138 1 woodwork 0.425399 0 1 inf
5141 1 wording 0.652778 0 1 inf
5143 1 worka NaN 0 1 inf
5145 1 workmanship 0.265836 0 1 inf
5146 1 worldly 0.031798 0 1 inf
5147 1 worm 0.652778 0 1 inf
5150 1 worthington NaN 0 1 inf
5153 1 wounds 0.079149 0 1 inf
5154 1 wracking 2.358220 0 1 inf
5155 1 wrapping 0.265836 0 1 inf
5157 1 wrath 0.031798 0 1 inf
5158 1 wriggly NaN 0 1 inf
5161 1 wrought 0.982790 0 1 inf
5162 1 yak NaN 0 1 inf
5163 1 yaks 2.358220 0 1 inf
5164 1 yank 0.154473 0 1 inf
5165 1 yanks 0.652778 0 1 inf
5172 1 yucatan 2.358220 0 1 inf
5173 1 zigzagging 0.652778 0 1 inf
5177 1 zoological NaN 0 1 inf
5178 1 zooming 2.358220 0 1 inf
\n", "

3934 rows \u00d7 6 columns

\n", "
" ], "metadata": {}, "output_type": "pyout", "prompt_number": 12, "text": [ " freq word log_likelihood brown_freq_normalized diff_absolute \\\n", "1 2437 a 39.837277 2125 312 \n", "2 1669 and -383.521998 2643 -974 \n", "4 1225 that 55.933699 970 255 \n", "6 1162 of -1776.113782 3336 -2174 \n", "8 934 in -616.702852 1954 -1020 \n", "9 921 to -1112.090217 2396 -1475 \n", "10 710 was -39.608135 899 -189 \n", "11 694 as 1.177540 664 30 \n", "12 667 it -22.394679 802 -135 \n", "13 585 for -97.381118 869 -284 \n", "14 566 you 165.609370 301 265 \n", "15 518 but 28.177335 401 117 \n", "17 471 i -0.008889 473 -2 \n", "18 471 is -254.363544 926 -455 \n", "19 442 so 234.694974 181 261 \n", "21 421 this -5.137938 471 -50 \n", "22 403 with -113.644883 667 -264 \n", "23 393 all 40.463187 274 119 \n", "24 384 on -94.773170 617 -233 \n", "25 383 his -112.694084 640 -257 \n", "26 364 gadsby NaN 0 364 \n", "28 327 up 96.888501 173 154 \n", "30 301 had -64.783953 470 -169 \n", "31 298 or -19.827346 385 -87 \n", "32 297 big 630.151174 32 265 \n", "33 291 not -42.458054 422 -131 \n", "34 289 an -8.174090 342 -53 \n", "35 284 at -96.568517 492 -208 \n", "36 280 out 31.832340 192 88 \n", "37 270 our 134.651348 114 156 \n", "... ... ... ... ... ... \n", "5122 1 wistaria NaN 0 1 \n", "5124 1 wit -0.421448 1 0 \n", "5126 1 withhold 1.487055 0 1 \n", "5128 1 wobbly 1.487055 0 1 \n", "5129 1 woild NaN 0 1 \n", "5130 1 wolf 0.265836 0 1 \n", "5133 1 wondrous 2.358220 0 1 \n", "5134 1 woo 0.982790 0 1 \n", "5137 1 woodlands NaN 0 1 \n", "5138 1 woodwork 0.425399 0 1 \n", "5141 1 wording 0.652778 0 1 \n", "5143 1 worka NaN 0 1 \n", "5145 1 workmanship 0.265836 0 1 \n", "5146 1 worldly 0.031798 0 1 \n", "5147 1 worm 0.652778 0 1 \n", "5150 1 worthington NaN 0 1 \n", "5153 1 wounds 0.079149 0 1 \n", "5154 1 wracking 2.358220 0 1 \n", "5155 1 wrapping 0.265836 0 1 \n", "5157 1 wrath 0.031798 0 1 \n", "5158 1 wriggly NaN 0 1 \n", "5161 1 wrought 0.982790 0 1 \n", "5162 1 yak NaN 0 1 \n", "5163 1 yaks 2.358220 0 1 \n", "5164 1 yank 0.154473 0 1 \n", "5165 1 yanks 0.652778 0 1 \n", "5172 1 yucatan 2.358220 0 1 \n", "5173 1 zigzagging 0.652778 0 1 \n", "5177 1 zoological NaN 0 1 \n", "5178 1 zooming 2.358220 0 1 \n", "\n", " diff_relative \n", "1 1.146824 \n", "2 0.631479 \n", "4 1.262887 \n", "6 0.348321 \n", "8 0.477994 \n", "9 0.384391 \n", "10 0.789766 \n", "11 1.045181 \n", "12 0.831671 \n", "13 0.673188 \n", "14 1.880399 \n", "15 1.291771 \n", "17 0.995772 \n", "18 0.508639 \n", "19 2.441989 \n", "21 0.893843 \n", "22 0.604198 \n", "23 1.434307 \n", "24 0.622366 \n", "25 0.598437 \n", "26 inf \n", "28 1.890173 \n", "30 0.640426 \n", "31 0.774026 \n", "32 9.281250 \n", "33 0.689573 \n", "34 0.845029 \n", "35 0.577236 \n", "36 1.458333 \n", "37 2.368421 \n", "... ... \n", "5122 inf \n", "5124 1.000000 \n", "5126 inf \n", "5128 inf \n", "5129 inf \n", "5130 inf \n", "5133 inf \n", "5134 inf \n", "5137 inf \n", "5138 inf \n", "5141 inf \n", "5143 inf \n", "5145 inf \n", "5146 inf \n", "5147 inf \n", "5150 inf \n", "5153 inf \n", "5154 inf \n", "5155 inf \n", "5157 inf \n", "5158 inf \n", "5161 inf \n", "5162 inf \n", "5163 inf \n", "5164 inf \n", "5165 inf \n", "5172 inf \n", "5173 inf \n", "5177 inf \n", "5178 inf \n", "\n", "[3934 rows x 6 columns]" ] } ], "prompt_number": 12 }, { "cell_type": "code", "collapsed": false, "input": [ "from matplotlib import pyplot as plt\n", "%matplotlib inline\n", "import numpy as np\n", "\n", "# x1_samples.shape -> (100, 2), 100 rows, 2 columns\n", "\n", "plt.figure(figsize=(8,6))\n", "\n", "dftemp = df[['word', 'log_likelihood', 'pct_the', 'freq']].dropna()\n", "#dftemp = dftemp[(dftemp.log_likelihood < -100) | (dftemp.log_likelihood > 100)]\n", "\n", "plt.scatter(dftemp.log_likelihood, dftemp.pct_the, s=dftemp.freq, marker='o', \n", " color='blue', alpha=0.4, label='data')\n", "#plt.scatter(x2_samples[:,0], x1_samples[:,1], marker='o', \n", " # color='green', alpha=0.7, label='x2 samples')\n", "#plt.scatter(x3_samples[:,0], x1_samples[:,1], marker='^', \n", " # color='red', alpha=0.7, label='x3 samples')\n", "plt.title(' ')\n", "plt.xlabel('Log-Likelihood that word is overrepresented in Gadsby')\n", "plt.ylabel(\"Percent probability of neighboring 'the' in Brown corpus\")" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "pyout", "prompt_number": 17, "text": [ "" ] }, { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAgEAAAGCCAYAAAB5KhCKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n", "text": [ "" ] } ], "prompt_number": 17 }, { "cell_type": "code", "collapsed": false, "input": [], "language": "python", "metadata": {}, "outputs": [] } ], "metadata": {} } ] }