Classifier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 528,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import collections\n",
    "import os\n",
    "import re\n",
    "import sys\n",
    "import math\n",
    "\n",
    "from nltk.stem.lancaster import LancasterStemmer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 529,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "'''\n",
    "Loads the Congressional Speech Data with labels\n",
    "@param root_dir Root directory of the files containing the speech text data\n",
    "Returns [(unigrams, bigrams, outcome=0/1)...]\n",
    "'''\n",
    "\n",
    "def load_raw_data(root_dir, stop_word_file, skip_label=False):\n",
    "    result = []\n",
    "    stop_words = set([])\n",
    "    # Any shorter implies insufficient data\n",
    "    MIN_WORDS = 5\n",
    "    st = LancasterStemmer()\n",
    "\n",
    "    if stop_word_file != None:\n",
    "        with open(stop_word_file) as f:\n",
    "            stop_words = set([line.strip() for line in f.readlines()])\n",
    "    \n",
    "    for file_name in os.listdir(root_dir):\n",
    "        if file_name.endswith(\".txt\"):\n",
    "            file_path = os.path.join(root_dir, file_name)\n",
    "            txt = None\n",
    "            \n",
    "            with open(file_path) as f:\n",
    "                txt = f.read()\n",
    "            \n",
    "            raw_words = []\n",
    "            \n",
    "            # A little cleanup for punctuations and case \n",
    "            for unclean_word in filter(None, re.split(\"[\\]\\[\\)\\(\\\"\\'\\n,\\-\\.!?: ]+\", txt)):\n",
    "                word = str.lower(unclean_word.strip())\n",
    "                # Random codecs can mess this\n",
    "                if len(word) > 2:\n",
    "                    raw_words.append(word)\n",
    "                \n",
    "            # Skip comments that are too short\n",
    "            if len(raw_words) < MIN_WORDS:\n",
    "                continue\n",
    "                \n",
    "            # Unigrams\n",
    "            unigrams = [word for word in raw_words if word not in stop_words]\n",
    "            bigrams = [\"%s %s\" % (raw_words[i], raw_words[i+1]) for i in range(len(raw_words) - 1)]\n",
    "            trigrams = [\"%s %s %s\" % (raw_words[i], raw_words[i+1], raw_words[i+2]) for i in range(len(raw_words) - 2)]\n",
    "            words = unigrams + bigrams + trigrams\n",
    "\n",
    "            yt = 0\n",
    "\n",
    "            # Outcome\n",
    "            # If skip_label is True, all labels are 0.\n",
    "            if not skip_label:\n",
    "                label = file_name.split(\".\")[0][-1]\n",
    "\n",
    "                if label == 'Y':\n",
    "                    yt = 1\n",
    "\n",
    "            result.append((words, yt, file_path))\n",
    "            \n",
    "    return result\n",
    "\n",
    "'''\n",
    "Returns\n",
    "X: [[unigrams + bigrams in review],...]\n",
    "y: outcomes [0,1,1,0...]\n",
    "cache: { \"word\" : [#0s where it appears, #1s where it appears]}\n",
    "'''\n",
    "def load_data(root_dir, stop_word_file=None, extract_names=False, print_every=100):\n",
    "    raw_data = load_raw_data(root_dir, stop_word_file)\n",
    "    X = [[] for rd in raw_data]\n",
    "    y = [0 for rd in raw_data]\n",
    "    names = [(None, None) for rd in raw_data]\n",
    "    file_paths = [None for rd in raw_data]\n",
    "   \n",
    "    cache = {}\n",
    "    n = len(raw_data)\n",
    "    \n",
    "    for i, sample in enumerate(raw_data):\n",
    "        if i%print_every == 0:\n",
    "            print(\"Processing %d/%d\" % (i+1, n))\n",
    "            sys.stdout.flush()\n",
    "            \n",
    "        words, outcome, file_path = sample\n",
    "        \n",
    "        # Just a good approximation.\n",
    "        if extract_names:\n",
    "            names[i] = (words[0], words[1])\n",
    "            words = words[2:]\n",
    "            \n",
    "        X[i] = words\n",
    "        y[i] = outcome\n",
    "        file_paths[i] = file_path\n",
    "        \n",
    "        for word in words:\n",
    "            if word not in cache:\n",
    "                cache[word] = [0,0]\n",
    "            \n",
    "            cache[word][outcome] += 1\n",
    "    \n",
    "    return np.array(X), np.array(y), cache, np.array(names), file_paths"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 530,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "'''\n",
    "DATA_ROOT=\"/Users/vikesh/development/opt/convote_v1.1/data_stage_three/\"\n",
    "TRAIN_ROOT=os.path.join(DATA_ROOT, \"training_set\")\n",
    "DEV_ROOT=os.path.join(DATA_ROOT, \"development_set\")\n",
    "TEST_ROOT=os.path.join(DATA_ROOT, \"test_set\")\n",
    "STOP_WORDS_FILE=\"stop-word-list.txt\"\n",
    "'''\n",
    "DIR_ROOT=\"/Users/vikesh/development/opt\"\n",
    "MODEL_ROOT=os.path.join(DIR_ROOT, \"model\")\n",
    "TRAIN_ROOT=os.path.join(MODEL_ROOT, \"train\")\n",
    "DEV_ROOT=os.path.join(MODEL_ROOT, \"dev\")\n",
    "TEST_ROOT=os.path.join(MODEL_ROOT, \"test\")\n",
    "\n",
    "DATA_ROOT=os.path.join(DIR_ROOT, \"data\")\n",
    "FIRST_NAMES_ROOT=os.path.join(DIR_ROOT, \"names\")\n",
    "LAST_NAMES_FILE=os.path.join(DIR_ROOT, \"names_census\", \"app.txt\")\n",
    "STOP_WORDS_FILE=\"stop-word-list.txt\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 531,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading data...\n",
      "Processing 1/532\n",
      "Processing 101/532\n",
      "Processing 201/532\n",
      "Processing 301/532\n",
      "Processing 401/532\n",
      "Processing 501/532\n",
      "Processing 1/146\n",
      "Processing 101/146\n",
      "Processing 1/75\n",
      "Loaded all data.\n",
      "Num train samples: 532\n",
      "Num -ive samples: 201\n",
      "Num +ive samples: 331\n",
      "Vocab: 65247\n"
     ]
    }
   ],
   "source": [
    "print(\"Loading data...\")\n",
    "# No names in train/dev/test set\n",
    "X_train, y_train, cache_train, names_null, file_paths_train = load_data(TRAIN_ROOT, STOP_WORDS_FILE)\n",
    "X_dev, y_dev, cache_dev, names_null ,file_paths_dev = load_data(DEV_ROOT, STOP_WORDS_FILE)\n",
    "X_test, y_test, cache_test, names_null, file_paths_test = load_data(TEST_ROOT, STOP_WORDS_FILE)\n",
    "print(\"Loaded all data.\")\n",
    "\n",
    "assert len(X_train) == len(y_train)\n",
    "print(\"Num train samples: %d\" % len(X_train))\n",
    "print(\"Num -ive samples: %d\" % np.sum(y_train==0))\n",
    "print(\"Num +ive samples: %d\" % np.sum(y_train==1))\n",
    "print(\"Vocab: %d\" % len(cache_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 611,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 615,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def naive_bayes(X, y, cache):\n",
    "    V = len(cache)\n",
    "    K=1.0\n",
    "    cnts = [np.sum(y==0), np.sum(y==1)]\n",
    "    \n",
    "    P = collections.defaultdict(lambda  : [K/(cnts[0] + K*V), K/(cnts[1] + K*V)])\n",
    "    Pc = [cnts[0]/float(len(y)), cnts[1]/float(len(y))]\n",
    "    \n",
    "    for word, counts in cache.iteritems():\n",
    "        for i, count in enumerate(counts):\n",
    "            # With laplace 1 smoothing\n",
    "            P[word][i] = float(count + K) / (cnts[i] + K*V)\n",
    "            \n",
    "    return P, Pc\n",
    "\n",
    "def evaluate(x, P, Pc):\n",
    "    lnps = [0,0]\n",
    "    \n",
    "    for i in range(0, 2):\n",
    "        lnps[i] = math.log(Pc[i]) + np.sum(np.log([P[word][i] for word in x]))\n",
    "    \n",
    "    return np.argmax(lnps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 616,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "P, Pc = naive_bayes(X_train, y_train, cache_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 534,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def test(X, P, Pc):\n",
    "    # Compute dev accuracy\n",
    "    y_eval = []\n",
    "\n",
    "    for x_eval in X:\n",
    "        y_eval.append(evaluate(x_eval, P, Pc))\n",
    "\n",
    "    return np.array(y_eval)\n",
    "\n",
    "# Fill the cache based on eval\n",
    "def backfill(X, y):\n",
    "    cache = collections.defaultdict(lambda : [0,0])\n",
    "    for i, x in enumerate(X):\n",
    "        for word in x:\n",
    "            cache[word][y[i]] += 1\n",
    "            \n",
    "    return cache"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 535,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train set accuracy: 1.000000\n",
      "Dev set accuracy: 0.924658\n",
      "Test set accuracy: 0.906667\n"
     ]
    }
   ],
   "source": [
    "# Compute train accuracy\n",
    "print(\"Train set accuracy: %f\" % np.mean(test(X_train, P, Pc)==y_train))\n",
    "# Compute dev accuracy\n",
    "print(\"Dev set accuracy: %f\" % np.mean(test(X_dev, P, Pc)==y_dev))\n",
    "# Compute test accuracy\n",
    "print(\"Test set accuracy: %f\" % np.mean(test(X_test, P, Pc)==y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 536,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing 1/13697\n",
      "Processing 1001/13697\n",
      "Processing 2001/13697\n",
      "Processing 3001/13697\n",
      "Processing 4001/13697\n",
      "Processing 5001/13697\n",
      "Processing 6001/13697\n",
      "Processing 7001/13697\n",
      "Processing 8001/13697\n",
      "Processing 9001/13697\n",
      "Processing 10001/13697\n",
      "Processing 11001/13697\n",
      "Processing 12001/13697\n",
      "Processing 13001/13697\n",
      "Loaded real data. Total=13697\n"
     ]
    }
   ],
   "source": [
    "X_data, y_data, cache_data, names_data, file_paths_data = load_data(DATA_ROOT, STOP_WORDS_FILE, extract_names=True, print_every=1000)\n",
    "assert(len(X_data) == len(y_data))\n",
    "print(\"Loaded real data. Total=%d\" % len(y_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 537,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# This is the real y_data\n",
    "y_eval = test(X_data, P, Pc)\n",
    "# Backfill the cache to have proper counts\n",
    "cache_data = backfill(X_data, y_eval)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 538,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Scatch space\n",
    "def toFile(x):\n",
    "    tmp = \"/tmp/data.txt\"\n",
    "    with open(tmp, \"w\") as f:\n",
    "        f.writelines(x)\n",
    "            \n",
    "    print tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 539,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Load the US names database\n",
    "# Returns the cache of name to its occurence in the DB\n",
    "def load_first_names(names_root, start_year=None, end_year=None):\n",
    "    cache = collections.defaultdict(int)\n",
    "    \n",
    "    for file_name in os.listdir(names_root):\n",
    "        if file_name.endswith(\".txt\"):\n",
    "            file_path = os.path.join(names_root, file_name)\n",
    "            year = int(file_name.split(\".\")[0][-4:])\n",
    "            \n",
    "            if start_year != None and year < start_year:\n",
    "                continue\n",
    "            if end_year != None and year > end_year:\n",
    "                continue\n",
    "                \n",
    "            names = []\n",
    "            \n",
    "            with open(file_path) as f:\n",
    "                for line in f.readlines():\n",
    "                    name, sex, occ_cnt = line.strip().split(\",\")\n",
    "                    cache[str.lower(name)] += long(occ_cnt)\n",
    "            \n",
    "    \n",
    "    return cache\n",
    "\n",
    "def load_last_names(names_file):\n",
    "    cache = collections.defaultdict(int)\n",
    "    \n",
    "    with open(names_file, \"r\") as f:\n",
    "        for line in f.readlines():\n",
    "            if not line.startswith(\"#\"):\n",
    "                data = line.strip().split()\n",
    "                cache[str.lower(data[0])] = long(data[2])\n",
    "            \n",
    "    return cache"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 540,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 93837 first names\n",
      "Loaded 151671 last names\n"
     ]
    }
   ],
   "source": [
    "# Load data necessary to compute stats\n",
    "first_name_cache = load_first_names(FIRST_NAMES_ROOT, start_year=1900)\n",
    "last_name_cache = load_last_names(LAST_NAMES_FILE)\n",
    "print(\"Loaded %s first names\" % len(first_name_cache))\n",
    "print(\"Loaded %s last names\" % len(last_name_cache))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 542,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of YES=9733, NO=3964\n",
      "Percentage of YES=0.710594, NO=0.289406\n"
     ]
    }
   ],
   "source": [
    "# Stats: Percentage of Yays and Nays\n",
    "cnt_yes = np.sum(y_eval == 1)\n",
    "cnt_no = np.sum(y_eval == 0)\n",
    "\n",
    "print(\"Number of YES=%d, NO=%d\" % (cnt_yes, cnt_no))\n",
    "print(\"Percentage of YES=%f, NO=%f\" % (float(cnt_yes)/len(y_eval), float(cnt_no)/len(y_eval)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 543,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Source: http://www.bloomberg.com/visual-data/best-and-worst//most-common-in-china-surnames \n",
    "with open(os.path.join(MODEL_ROOT, \"chinese_names.txt\"), \"r\") as f:\n",
    "    TOP_ASIAN_NAMES = set([str.lower(line.strip()) for line in f.readlines()])\n",
    "    \n",
    "# Source: Wikipedia, and manually eyeballing the data\n",
    "with open(os.path.join(MODEL_ROOT, \"indian_names.txt\"), \"r\")  as f:\n",
    "    TOP_INDIAN_NAMES = set([str.lower(line.strip()) for line in f.readlines()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 544,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Stats: How many names (approximately) are from US citizens/residents\n",
    "def get_amr(names, first_name_cache, last_name_cache, percentile, exclude_intl_names=True):\n",
    "    min_occ_first = np.percentile([occ_cnt for name, occ_cnt in first_name_cache.iteritems()], percentile)\n",
    "    min_occ_last = np.percentile([occ_cnt for name, occ_cnt in last_name_cache.iteritems()], percentile)\n",
    "    \n",
    "    amr = np.zeros(len(names))\n",
    "    \n",
    "    for i, name in enumerate(names):\n",
    "        # This is a good approximation\n",
    "        try:\n",
    "            first_name, last_name = name\n",
    "            # Probability of US name\n",
    "            is_name_us = (first_name_cache[first_name] > min_occ_first and last_name_cache[last_name] > min_occ_last)\n",
    "            is_name_not_intl = last_name not in TOP_ASIAN_NAMES and last_name not in TOP_INDIAN_NAMES\n",
    "            amr[i] = is_name_us and (not exclude_intl_names or is_name_not_intl)\n",
    "        except:\n",
    "            print(\"WARNING: Bad data. Skipping.\")\n",
    "            \n",
    "    return amr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 597,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Excluding intl names: False\n",
      "# Percentile\tUS ratio\tUS among +ve\tUS among -ve\t+ve among US\t-ve among US\n",
      "60.000\t0.495\t0.386\t0.763\t0.554\t0.446\n",
      "75.000\t0.450\t0.337\t0.726\t0.533\t0.467\n",
      "80.000\t0.423\t0.309\t0.703\t0.519\t0.481\n",
      "85.000\t0.396\t0.280\t0.681\t0.503\t0.497\n",
      "88.000\t0.374\t0.260\t0.653\t0.494\t0.506\n",
      "90.000\t0.358\t0.245\t0.633\t0.488\t0.512\n",
      "91.000\t0.350\t0.238\t0.624\t0.484\t0.516\n",
      "92.000\t0.340\t0.231\t0.607\t0.482\t0.518\n",
      "93.000\t0.329\t0.221\t0.595\t0.477\t0.523\n",
      "94.000\t0.320\t0.214\t0.581\t0.475\t0.525\n",
      "95.000\t0.308\t0.206\t0.559\t0.475\t0.525\n",
      "96.000\t0.288\t0.189\t0.530\t0.468\t0.532\n",
      "97.000\t0.265\t0.172\t0.494\t0.461\t0.539\n",
      "98.000\t0.234\t0.150\t0.441\t0.456\t0.544\n",
      "99.000\t0.186\t0.120\t0.348\t0.458\t0.542\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAi0AAAEPCAYAAACOZfTSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10fb7f2d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Excluding intl names: True\n",
      "# Percentile\tUS ratio\tUS among +ve\tUS among -ve\t+ve among US\t-ve among US\n",
      "60.000\t0.363\t0.202\t0.760\t0.394\t0.606\n",
      "75.000\t0.341\t0.185\t0.723\t0.386\t0.614\n",
      "80.000\t0.328\t0.176\t0.700\t0.382\t0.618\n",
      "85.000\t0.316\t0.169\t0.678\t0.380\t0.620\n",
      "88.000\t0.303\t0.162\t0.650\t0.380\t0.620\n",
      "90.000\t0.293\t0.155\t0.631\t0.377\t0.623\n",
      "91.000\t0.288\t0.152\t0.622\t0.375\t0.625\n",
      "92.000\t0.280\t0.148\t0.605\t0.375\t0.625\n",
      "93.000\t0.273\t0.144\t0.592\t0.373\t0.627\n",
      "94.000\t0.267\t0.139\t0.579\t0.372\t0.628\n",
      "95.000\t0.257\t0.134\t0.557\t0.372\t0.628\n",
      "96.000\t0.243\t0.126\t0.528\t0.370\t0.630\n",
      "97.000\t0.225\t0.117\t0.492\t0.368\t0.632\n",
      "98.000\t0.202\t0.105\t0.440\t0.371\t0.629\n",
      "99.000\t0.160\t0.084\t0.347\t0.373\t0.627\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAi0AAAEPCAYAAACOZfTSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10fb7f8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "PERCENTILES = [60, 75, 80, 85, 88, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99]\n",
    "\n",
    "for exclude_intl_names in [False, True]:\n",
    "    print(\"# Excluding intl names: {}\".format(exclude_intl_names))\n",
    "    print(\"# Percentile\\tUS ratio\\tUS among +ve\\tUS among -ve\\t+ve among US\\t-ve among US\")\n",
    "    plt_percentiles = []\n",
    "    plt_US_pos = []\n",
    "    plt_US_neg = []\n",
    "    plt_pos_US = []\n",
    "    plt_neg_US = []\n",
    "\n",
    "    for percentile in PERCENTILES:\n",
    "        amr = get_amr(names_data, first_name_cache, last_name_cache, percentile, exclude_intl_names)\n",
    "\n",
    "        with open(\"results/perc.%f.txt\" % percentile, \"w\") as f:\n",
    "            for i, name in enumerate(names_data):\n",
    "                if amr[i] == 1:\n",
    "                    f.write(str(amr[i]) + \", \" + str(names_data[i]) + \", VOTE=\" + str(y_eval[i]) + \"\\n\")\n",
    "                    f.write(\" \".join(X_data[i]) + \"\\n\")\n",
    "                    f.write(\"*******\\n\")\n",
    "\n",
    "        plt_percentiles.append(percentile)\n",
    "        plt_US_pos.append(np.sum(y_eval[amr==1])/float(np.sum(y_eval==1)))\n",
    "        plt_US_neg.append(np.sum((1-y_eval)[amr==1])/float(np.sum(1-y_eval)))\n",
    "        plt_pos_US.append(np.sum(y_eval[amr==1])/float(np.sum(amr)))\n",
    "        plt_neg_US.append(np.sum((1-y_eval)[amr==1])/float(np.sum(amr==1)))\n",
    "\n",
    "        print(\"{0:0.3f}\\t{1:0.3f}\\t{2:0.3f}\\t{3:0.3f}\\t{4:0.3f}\\t{5:0.3f}\".format(\n",
    "                percentile,\n",
    "                np.sum(amr==1)/float(len(y_eval)),\n",
    "                plt_US_pos[-1],\n",
    "                plt_US_neg[-1],\n",
    "                plt_pos_US[-1],\n",
    "                plt_neg_US[-1],\n",
    "            )\n",
    "         )\n",
    "\n",
    "    plt.xlabel(\"Percentile\")\n",
    "    plt.ylabel(\"Values\")\n",
    "    plt.plot(np.array(plt_percentiles), np.array(plt_US_pos), label=\"US among positives\")\n",
    "    plt.plot(np.array(plt_percentiles), np.array(plt_US_neg), label=\"US among negatives\")\n",
    "    plt.plot(np.array(plt_percentiles), np.array(plt_pos_US), label=\"Postitives among US\")\n",
    "    plt.plot(np.array(plt_percentiles), np.array(plt_neg_US), label=\"Negatives among US\")\n",
    "\n",
    "    plt.legend(loc='center left', bbox_to_anchor=(1, .5))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 547,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "min_first = np.percentile([occ for name, occ in first_name_cache.iteritems()], percentile)\n",
    "min_last = np.percentile([occ for name, occ in last_name_cache.iteritems()], percentile)\n",
    "\n",
    "with open(\"results/neg.not.us.txt\", \"w\") as f:\n",
    "    for i, name in enumerate(names_data):\n",
    "        if amr[i] == 0 and y_eval[i] == 0:\n",
    "            first_name, last_name = name\n",
    "            f.write(\"NAME=\" + str(name) + \", SCORE=%s/%s, %s/%s\" % (first_name_cache[first_name], min_first, last_name_cache[last_name], min_last))\n",
    "            f.write(\"\\n\" + \" \".join(X_data[i]))\n",
    "            f.write(\"\\n\")\n",
    "            f.write(\"*******\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 548,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "min_first = np.percentile([occ for name, occ in first_name_cache.iteritems()], percentile)\n",
    "min_last = np.percentile([occ for name, occ in last_name_cache.iteritems()], percentile)\n",
    "\n",
    "with open(\"results/pos.us.txt\", \"w\") as f:\n",
    "    for i, name in enumerate(names_data):\n",
    "        if amr[i] == 1 and y_eval[i] == 1:\n",
    "            first_name, last_name = name\n",
    "            f.write(\"NAME=\" + str(name) + \", SCORE=%s/%s, %s/%s\" % (first_name_cache[first_name], min_first, last_name_cache[last_name], min_last))\n",
    "            f.write(\"\\n\" + \" \".join(X_data[i]))\n",
    "            f.write(\"\\n\")\n",
    "            f.write(\"*******\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 592,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Stats: Most common words\n",
    "def get_unigrams(arr):\n",
    "    return [word for word in arr if \" \" not in word]\n",
    "\n",
    "# cond is a np array of truth values == len(y)\n",
    "def get_word_freq(X, cond):\n",
    "    freqs_cache = collections.defaultdict(int)\n",
    "    \n",
    "    for x in X[cond]:\n",
    "        for unigram in get_unigrams(x):\n",
    "            freqs_cache[unigram] += 1\n",
    "            \n",
    "    return sorted([(word, count) for word, count in freqs_cache.iteritems()], reverse=True, key=lambda x: x[1])\n",
    "\n",
    "# Get the rank (index) of word in the sorted freqs array\n",
    "def get_ranks(freqs):\n",
    "    return {pair[0]:i for i, pair in enumerate(freqs)}\n",
    "\n",
    "# Write the file for creating word clouds\n",
    "def write_cloud(freqs_k, file_path):\n",
    "    fstr = \"\"\n",
    "    for tup in freqs_k:\n",
    "        word, freq = tup\n",
    "        # print word, freq\n",
    "        for i in range(freq):\n",
    "            fstr += \"%s \" % word\n",
    "\n",
    "    with open(file_path, \"w\") as f:\n",
    "        f.write(fstr)\n",
    "    \n",
    "    print(\"Written to file: {}\".format(file_path))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 589,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Most common words by americans\n",
    "TOP_K = 50\n",
    "    \n",
    "amr = get_amr(names_data, first_name_cache, last_name_cache, percentile=99)\n",
    "# switch y_eval\n",
    "freqs = get_word_freq(X_data, (y_eval==0) * (amr==1))\n",
    "\n",
    "for tup in freqs[:TOP_K]:\n",
    "    word, freq = tup\n",
    "    # print word, freq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 590,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Most common words in general\n",
    "amr = get_amr(names_data, first_name_cache, last_name_cache, percentile=99)\n",
    "suffixes = [\"neg\", \"pos\"]\n",
    "TOP_K = 50\n",
    "\n",
    "for label in range(0, 2):\n",
    "    # switch y_eval\n",
    "    freqs = get_word_freq(X_data, y_eval==label)\n",
    "    fstr = \"\"\n",
    "\n",
    "    with open(\"results/words.{}.txt\".format(suffixes[label]), \"w\") as f:\n",
    "        total = np.sum([pair[1] for pair in freqs])\n",
    "        f.writelines(map(lambda x: \"%s\\t%d\\t%f\\n\" % (x[0], x[1], (x[1]*100)/float(total)), freqs))\n",
    "        \n",
    "    write_cloud(freqs[:TOP_K], \"results/cloud.{}.txt\".format(suffixes[label]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 552,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBwAAAJyCAYAAABntdMVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCRwAAACAxQkcAAAAgMUJHAAAAIDF/Q7UPhjF7yhf5gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10f789350>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Bar chart for word usage\n",
    "freqs = get_word_freq(X_data, y_eval < 2)\n",
    "freqs_pos = get_word_freq(X_data, y_eval == 1)\n",
    "ranks_pos = get_ranks(freqs_pos)\n",
    "\n",
    "freqs_neg = get_word_freq(X_data, y_eval == 0)\n",
    "ranks_neg = get_ranks(freqs_neg)\n",
    "\n",
    "total_neg = np.sum([cnt for word, cnt in freqs_neg])\n",
    "total_pos = np.sum([cnt for word, cnt in freqs_pos])\n",
    "\n",
    "TOP_K = 20\n",
    "np_words = []\n",
    "np_freq_pos = []\n",
    "np_freq_neg = []\n",
    "np_rank_pos = []\n",
    "np_rank_neg = []\n",
    "\n",
    "for pair in freqs[:TOP_K]:\n",
    "    word, count = pair\n",
    "    np_words.append(word)\n",
    "    np_freq_neg.append((cache_data[word][0]*100)/float(total_neg))\n",
    "    np_freq_pos.append((cache_data[word][1]*100)/float(total_pos))\n",
    "    \n",
    "    np_rank_pos.append(ranks_pos[word])\n",
    "    np_rank_neg.append(ranks_neg[word])\n",
    "    \n",
    "# Plot this.\n",
    "index = np.arange(0, TOP_K*2, 2)\n",
    "bar_width = 0.55\n",
    "\n",
    "plt.figure(figsize=(18,10))\n",
    "\n",
    "rects1 = plt.bar(index, np.array(np_freq_neg), bar_width,\n",
    "                 color='r',\n",
    "                 label='% occurence in -ve comments')\n",
    "\n",
    "rects2 = plt.bar(index + bar_width, np.array(np_freq_pos), bar_width,\n",
    "                 color='b',\n",
    "                 label='% occurence in +ve comments')\n",
    "\n",
    "\n",
    "plt.xticks(index + bar_width, np_words)\n",
    "\n",
    "plt.xlabel('Word')\n",
    "plt.ylabel('Occurence')\n",
    "plt.title('Word occurences in +ve and -ve reviews')\n",
    "plt.legend()\n",
    "\n",
    "def autolabel(rects, ranks):\n",
    "    # attach some text labels\n",
    "    for i, rect in enumerate(rects):\n",
    "        word, count = freqs[i]\n",
    "        height = rect.get_height()\n",
    "        plt.text(rect.get_x() + rect.get_width()/2., 1.05*height,\n",
    "                '%d' % (ranks[word] + 1),\n",
    "                ha='center', va='bottom')\n",
    "\n",
    "autolabel(rects1, ranks_neg)\n",
    "autolabel(rects2, ranks_pos)\n",
    "\n",
    "# plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 585,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtsAAAH2CAYAAAChqkDdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZKUE8O2JEmSlJP/D+6XdTFHV0qRAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11c0ee550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Confirm the Zipf's law\n",
    "freqs = get_word_freq(X_data, y_eval < 2)\n",
    "np_occ = np.array([cnt for word, cnt in freqs])\n",
    "np_rank = np.array([i for i, pair in enumerate(freqs)])\n",
    "\n",
    "plt.figure(figsize=(12, 8))\n",
    "plt.yscale('log')\n",
    "plt.xscale('log')\n",
    "plt.xlabel(\"Rank\")\n",
    "plt.ylabel(\"Frequency of occurence\")\n",
    "plt.plot(np_rank, np_occ)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 599,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "artificially 243\n",
      "65% 242\n",
      "2013; 173\n",
      "568 123\n",
      "grassley 106\n",
      "administratively 106\n",
      "susceptible 104\n",
      "takeaways 72\n",
      "mayorkas 49\n",
      "concurred 49\n",
      "elders 48\n",
      "defining 45\n",
      "cio 41\n",
      "unionize 40\n",
      "exempts 38\n",
      "categorizes 37\n",
      "hewlett 35\n",
      "norm 34\n",
      "packard 34\n",
      "sce 33\n",
      "suppression 31\n",
      "afl 27\n",
      "yearsthe 25\n",
      "1b} 25\n",
      "validthere 24\n",
      "crony 23\n",
      "immobile 22\n",
      "juan 22\n",
      "sergei 21\n",
      "penalty 21\n",
      "poached 21\n",
      "274 17\n",
      "penalties 17\n",
      "497 17\n",
      "howard 16\n",
      "visasbut 14\n",
      "undercutting 13\n",
      "unamerican 12\n",
      "donors 11\n",
      "illusionary 11\n",
      "betrayal 11\n",
      "invaders 10\n",
      "rolley 10\n",
      "dictator 10\n",
      "alf 10\n",
      "foia 9\n",
      "punishment 8\n",
      "adamantly 8\n",
      "robs 8\n",
      "gentlemen 8\n",
      "Written to file: results/cloud.exc.neg.txt\n"
     ]
    }
   ],
   "source": [
    "# Stats: Words in -ve reviews that are not in +ve reviews\n",
    "TOP_K = 50\n",
    "\n",
    "freqs_pos = get_word_freq(X_data, y_eval == 1)\n",
    "freqs_neg = get_word_freq(X_data, y_eval == 0)\n",
    "\n",
    "words_pos = set([word for word, cnt in freqs_pos])\n",
    "words_neg = set([word for word, cnt in freqs_neg])\n",
    "\n",
    "ranks_pos = get_ranks(freqs_pos)\n",
    "ranks_neg = get_ranks(freqs_neg)\n",
    "\n",
    "words_exc_neg = words_neg - words_pos\n",
    "freqs_exc_neg = []\n",
    "\n",
    "for word_exc_neg in words_exc_neg:\n",
    "    freqs_exc_neg.append(freqs_neg[ranks_neg[word_exc_neg]])\n",
    "\n",
    "freqs_exc_neg = sorted(freqs_exc_neg, reverse=True, key=lambda x: x[1])\n",
    "\n",
    "for word_exc, count_exc in freqs_exc_neg[:TOP_K]:\n",
    "    print word_exc, count_exc\n",
    "\n",
    "write_cloud(freqs_exc_neg[:TOP_K], \"results/cloud.exc.neg.txt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 556,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of +ve among lost: 0.466735966736\n"
     ]
    }
   ],
   "source": [
    "# Stat: Who did we lose from 60 percentile to 90 percentile and what were their names and comments\n",
    "LOW_PERC = 60\n",
    "HIGH_PERCS = [65, 70, 75, 80, 85, 90, 95, 99]\n",
    "\n",
    "for high_perc in [90]:\n",
    "    low_amr = get_amr(names_data, first_name_cache, last_name_cache, LOW_PERC, exclude_intl_names=True)\n",
    "    high_amr = get_amr(names_data, first_name_cache, last_name_cache, high_perc, exclude_intl_names=True)\n",
    "\n",
    "    lost_idx = (1-high_amr) * low_amr\n",
    "    lost_votes = y_eval[lost_idx == 1]\n",
    "    print(\"Percentage of +ve among lost: {}\".format(np.sum(lost_votes) / float(len(lost_votes))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 584,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LABEL=0\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCkAAAJPCAYAAABRi44aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABoBYIUAAAAAACgFQhSAAAAAACAVvh/K5p1k5A4mEIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1149b0710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LABEL=1\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCIAAAJPCAYAAACpa3XtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1260dee10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Stat: Most common names \n",
    "def get_top_names(names, cond):\n",
    "    names = names_data[cond]\n",
    "    last_names_cnt_cache = collections.defaultdict(int)\n",
    "    \n",
    "    for first_name, last_name in names:\n",
    "        last_names_cnt_cache[last_name] += 1\n",
    "        \n",
    "    return sorted([(last_name, float(cnt)/len(names)) for last_name, cnt in last_names_cnt_cache.iteritems()], \n",
    "                  reverse=True, key=lambda x: x[1])\n",
    "\n",
    "labels = [0,1]\n",
    "TOP_K = 20\n",
    "index = np.arange(0, TOP_K*2, 2)\n",
    "\n",
    "for label in labels:\n",
    "    print(\"LABEL=%d\" % label)\n",
    "    names = get_top_names(names_data, y_eval == label)\n",
    "    np_names, np_freq = zip(*names[:TOP_K])\n",
    "    \n",
    "    plt.figure(figsize=(18,10))\n",
    "    bar_width = 0.70\n",
    "    \n",
    "    rects1 = plt.bar(index, np.array(np_freq), bar_width,\n",
    "                 color='r', label=\"Frequency in %s-labeled comments\" % label)\n",
    "                 \n",
    "    plt.xticks(index + bar_width, np_names)\n",
    "    plt.ylabel(\"Frequency of occurence of names\")\n",
    "    plt.legend()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 619,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import dill\n",
    "dill.dump(P, open(\"model/P.p\", \"wb\" ))\n",
    "dill.dump(Pc, open(\"model/Pc.p\", \"wb\" ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}