Word Embeddings SandBox.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Word Embeddings For Text Classification\n",
    "Source: https://github.com/RaRe-Technologies/movie-plots-by-genre/blob/master/Document%20classification%20with%20word%20embeddings%20tutorial.ipynb\n",
    "\n",
    "Slides with Notes: https://www.slideshare.net/YiShinChen1/quick-tour-of-text-mining"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TASK: \n",
    "\n",
    "Classify Movie plots by genre using word embeddings techniques in Gensim"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REPO:\n",
    "https://goo.gl/ppHX65"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Requirements:\n",
    "\n",
    "#### Software:\n",
    "------------------\n",
    "\n",
    "* Python 3 (coding will be done strictly using Python)\n",
    "* Anaconda Environment (recommended but not mandatory) (https://www.continuum.io/downloads)\n",
    "* Jupyter (http://jupyter.org/)\n",
    "* Google's word2vec (Download the file... warning! it is really huge)(https://drive.google.com/file/d/0B7XkCwpI5KDYNlNUTTlSS21pQmM/edit?usp=sharing)\n",
    "* Gensim (https://radimrehurek.com/gensim/)\n",
    "* Scikit Learn (http://scikit-learn.org/stable/) (get the latest version)\n",
    "* Pandas (http://pandas.pydata.org/)\n",
    "* Matplotlib (https://matplotlib.org/)\n",
    "* NLTK (for stopwords) (http://www.nltk.org/)\n",
    "\n",
    "#### Computing Resources:\n",
    "-------------------\n",
    "* Operating System: Preferably Linux or MacOS (Windows break but you can try it out)\n",
    "* RAM: 4GB \n",
    "* Disk Space: 8GB (mostly to store word embeddings)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "import logging\n",
    "logging.root.handlers = []  # Jupyter messes up logging so needs a reset\n",
    "logging.basicConfig(format='%(asctime)s : %(levelname)s : %(message)s', level=logging.INFO)\n",
    "from smart_open import smart_open\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from numpy import random\n",
    "import gensim\n",
    "import nltk\n",
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn import linear_model\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "import matplotlib.pyplot as plt\n",
    "from gensim.models import Word2Vec\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn import linear_model\n",
    "from nltk.corpus import stopwords\n",
    "from helpers import *\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Exploration\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "171156"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# read reviews\n",
    "df = pd.read_csv('data/movie_reviews/tagged_plots_movielens.csv')\n",
    "df = df.dropna()\n",
    "\n",
    "# count the number of words\n",
    "df['plot'].apply(lambda x: len(x.split(' '))).sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Plot tokens by frequency"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note:** Usually at least 500K words are suggested to train word2vec model, so performance may be unsual here\n",
    "\n",
    "We only have (1/5th) therefore some models may not peform very well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>movieId</th>\n",
       "      <th>plot</th>\n",
       "      <th>tag</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>A little boy named Andy loves to be in his roo...</td>\n",
       "      <td>animation</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>When two kids find and play a magical board ga...</td>\n",
       "      <td>fantasy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>Things don't seem to change much in Wabasha Co...</td>\n",
       "      <td>comedy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>Hunters and their prey--Neil and his professio...</td>\n",
       "      <td>action</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>An ugly duckling having undergone a remarkable...</td>\n",
       "      <td>romance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>Some terrorists kidnap the Vice President of t...</td>\n",
       "      <td>action</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>10</td>\n",
       "      <td>James Bond teams up with the lone survivor of ...</td>\n",
       "      <td>action</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>15</td>\n",
       "      <td>Morgan Adams and her slave, William Shaw, are ...</td>\n",
       "      <td>action</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>17</td>\n",
       "      <td>When Mr. Dashwood dies, he must leave the bulk...</td>\n",
       "      <td>romance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>18</td>\n",
       "      <td>This movie features the collaborative director...</td>\n",
       "      <td>comedy</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  movieId                                               plot  \\\n",
       "0           0        1  A little boy named Andy loves to be in his roo...   \n",
       "1           1        2  When two kids find and play a magical board ga...   \n",
       "2           2        3  Things don't seem to change much in Wabasha Co...   \n",
       "3           3        6  Hunters and their prey--Neil and his professio...   \n",
       "4           4        7  An ugly duckling having undergone a remarkable...   \n",
       "5           5        9  Some terrorists kidnap the Vice President of t...   \n",
       "6           6       10  James Bond teams up with the lone survivor of ...   \n",
       "7           7       15  Morgan Adams and her slave, William Shaw, are ...   \n",
       "8           8       17  When Mr. Dashwood dies, he must leave the bulk...   \n",
       "9           9       18  This movie features the collaborative director...   \n",
       "\n",
       "         tag  \n",
       "0  animation  \n",
       "1    fantasy  \n",
       "2     comedy  \n",
       "3     action  \n",
       "4    romance  \n",
       "5     action  \n",
       "6     action  \n",
       "7     action  \n",
       "8    romance  \n",
       "9     comedy  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f9eadcc4e80>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9eadcb1dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# classes\n",
    "my_tags = ['sci-fi' , 'action', 'comedy', 'fantasy', 'animation', 'romance']\n",
    "df.tag.value_counts().plot(kind=\"bar\", rot=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note:** Unbalanced dataset; Comedy has significantly more examples than the rest of the classes (40% overall)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**TODO:** Do some more text mining exploration and statistical inferences here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# split the data (90/10)\n",
    "train_data, test_data = train_test_split(df, test_size=0.1, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(243, 2184)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(test_data), len(train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f9ead443828>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD8CAYAAABn919SAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9eadcc2f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# distribution of the test data\n",
    "test_data.tag.value_counts().plot(kind=\"bar\", rot=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train Naive Models (Baselines)\n",
    "---\n",
    "* Bag of words\n",
    "* N-grams\n",
    "* TF-IDF"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Bag of Words\n",
    "Using scikitlearn CountVectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# training\n",
    "count_vectorizer = CountVectorizer(\n",
    "    analyzer=\"word\", tokenizer=nltk.word_tokenize,\n",
    "    preprocessor=None, stop_words='english', max_features=3000) \n",
    "\n",
    "# choose what are the features\n",
    "train_data_features = count_vectorizer.fit_transform(train_data['plot'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise**: Authors only use logistic regression but we can also train with naive bayesian, random forest, gradient boosting, and deep neural networks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "logreg_model = linear_model.LogisticRegression(n_jobs=1, C=1e5)\n",
    "logreg_model = logreg_model.fit(train_data_features, train_data['tag'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['warrior',\n",
       " 'warriors',\n",
       " 'wars',\n",
       " 'washington',\n",
       " 'watch',\n",
       " 'watches',\n",
       " 'watching',\n",
       " 'water',\n",
       " 'waters',\n",
       " 'way',\n",
       " 'wayne']"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# oberve some features\n",
    "count_vectorizer.get_feature_names()[2899:2910]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1x3000 sparse matrix of type '<class 'numpy.int64'>'\n",
       "\twith 43 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# amount of features detected (3000)\n",
    "train_data_features[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.423868312757\n",
      "confusion matrix\n",
      " [[21  2 10  1  4  4]\n",
      " [ 4 10  8  0  3  6]\n",
      " [10 11 45  3 16  1]\n",
      " [ 1  5  3  4  3  0]\n",
      " [ 1  3 15  2 11  3]\n",
      " [ 8  5  6  2  0 12]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAEmCAYAAADfpHMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9ead8cd208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.predict(count_vectorizer, logreg_model, test_data, my_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note 1**: Classifier performs slighly better than chance. 40% if the classifier said all were comedy and 42% with the bag of word model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# words for the action genre\n",
    "comedy_tag_id = word_embeddings.get_tag_index(my_tags, \"comedy\")\n",
    "comedy_words = word_embeddings.most_influential_words(logreg_model, count_vectorizer, \\\n",
    "                                                      comedy_tag_id, 3000)\n",
    "comedy_words = pd.DataFrame(comedy_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>mistaken</td>\n",
       "      <td>11.810842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>jewish</td>\n",
       "      <td>10.668508</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>suspects</td>\n",
       "      <td>10.104922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>comedy</td>\n",
       "      <td>9.949616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>dealer</td>\n",
       "      <td>9.775865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>comedian</td>\n",
       "      <td>9.634339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>operation</td>\n",
       "      <td>9.512068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>stuart</td>\n",
       "      <td>9.323364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>actress</td>\n",
       "      <td>9.232942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>dimension</td>\n",
       "      <td>9.097288</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0          1\n",
       "0   mistaken  11.810842\n",
       "1     jewish  10.668508\n",
       "2   suspects  10.104922\n",
       "3     comedy   9.949616\n",
       "4     dealer   9.775865\n",
       "5   comedian   9.634339\n",
       "6  operation   9.512068\n",
       "7     stuart   9.323364\n",
       "8    actress   9.232942\n",
       "9  dimension   9.097288"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comedy_words[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "# words for the action genre\n",
    "animation_tag_id = word_embeddings.get_tag_index(my_tags, \"animation\")\n",
    "animation_words = word_embeddings.most_influential_words(logreg_model, count_vectorizer,\\\n",
    "                                                         animation_tag_id, 3000)\n",
    "animation_words = pd.DataFrame(animation_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>nazi</td>\n",
       "      <td>8.946895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>sisters</td>\n",
       "      <td>8.870921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>troubled</td>\n",
       "      <td>8.720853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>affair</td>\n",
       "      <td>8.603380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>revealed</td>\n",
       "      <td>8.499657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationships</td>\n",
       "      <td>8.112722</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>decide</td>\n",
       "      <td>8.016373</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>spending</td>\n",
       "      <td>7.919799</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>wolf</td>\n",
       "      <td>7.470311</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>photographer</td>\n",
       "      <td>7.268472</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               0         1\n",
       "0           nazi  8.946895\n",
       "1        sisters  8.870921\n",
       "2       troubled  8.720853\n",
       "3         affair  8.603380\n",
       "4       revealed  8.499657\n",
       "5  relationships  8.112722\n",
       "6         decide  8.016373\n",
       "7       spending  7.919799\n",
       "8           wolf  7.470311\n",
       "9   photographer  7.268472"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# the most influential words for the animation category\n",
    "animation_words[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>national</td>\n",
       "      <td>15.013773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>chosen</td>\n",
       "      <td>7.682396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>princess</td>\n",
       "      <td>6.555631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>prove</td>\n",
       "      <td>6.072531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>beast</td>\n",
       "      <td>5.901359</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>fan</td>\n",
       "      <td>5.886863</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>moving</td>\n",
       "      <td>5.641193</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>fantasies</td>\n",
       "      <td>5.636695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>director</td>\n",
       "      <td>5.516928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>angel</td>\n",
       "      <td>5.485528</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0          1\n",
       "0   national  15.013773\n",
       "1     chosen   7.682396\n",
       "2   princess   6.555631\n",
       "3      prove   6.072531\n",
       "4      beast   5.901359\n",
       "5        fan   5.886863\n",
       "6     moving   5.641193\n",
       "7  fantasies   5.636695\n",
       "8   director   5.516928\n",
       "9      angel   5.485528"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# words for the action genre\n",
    "fantasy_tag_id = word_embeddings.get_tag_index(my_tags, \"fantasy\")\n",
    "fantasy_words = word_embeddings.most_influential_words(logreg_model, count_vectorizer,\\\n",
    "                                                       fantasy_tag_id, 3000)\n",
    "fantasy_words = pd.DataFrame(fantasy_words)\n",
    "\n",
    "# the most influential words for the animation category\n",
    "fantasy_words[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[['jay', 12],\n",
       "  ['insane', 114],\n",
       "  ['lawyer', 116],\n",
       "  ['environment', 162],\n",
       "  ['marry', 166],\n",
       "  ['did', 170],\n",
       "  ['sons', 222],\n",
       "  ['hunting', 302],\n",
       "  ['boys', 318],\n",
       "  ['asked', 323],\n",
       "  ['jobs', 336],\n",
       "  ['opportunity', 341],\n",
       "  ['heading', 408],\n",
       "  ['threatened', 414],\n",
       "  ['american', 427],\n",
       "  ['brooklyn', 428],\n",
       "  ['turned', 429],\n",
       "  ['late', 445],\n",
       "  ['chinese', 450],\n",
       "  ['president', 453],\n",
       "  ['ready', 456],\n",
       "  ['look', 469],\n",
       "  ['england', 488]],\n",
       " 23]"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check what are the words that overlap between comedy and animation\n",
    "word_embeddings.check_word_overlap(comedy_words[:500][0], animation_words[:500][0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Character N-grams"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A character n-gram is a chunk of a document of length n. It is a poor man's tokenizer but sometimes works well. The parameter n depends on language and the corpus. We choose length between 3 and 6 characters and to only focus on 3k most popular ones. 3K is a fair comparison since the previous bag of words model was this size as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_gram_vectorizer = CountVectorizer(\n",
    "    analyzer=\"char\",\n",
    "    ngram_range=([2,5]),\n",
    "    tokenizer=None,    \n",
    "    preprocessor=None,                               \n",
    "    max_features=3000) \n",
    "\n",
    "charn_model = linear_model.LogisticRegression(n_jobs=1, C=1e5)\n",
    "\n",
    "train_data_features = n_gram_vectorizer.fit_transform(train_data['plot'])\n",
    "\n",
    "charn_model = charn_model.fit(train_data_features, train_data['tag'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[' chil',\n",
       " ' ci',\n",
       " ' cit',\n",
       " ' cl',\n",
       " ' co',\n",
       " ' col',\n",
       " ' com',\n",
       " ' come',\n",
       " ' comp',\n",
       " ' con',\n",
       " ' cont',\n",
       " ' cou',\n",
       " ' cr',\n",
       " ' cre',\n",
       " ' cu',\n",
       " ' d',\n",
       " ' da',\n",
       " ' dan',\n",
       " ' day',\n",
       " ' de']"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# some features\n",
    "n_gram_vectorizer.get_feature_names()[100:120]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.399176954733\n",
      "confusion matrix\n",
      " [[17  3 12  1  5  4]\n",
      " [ 4  9  9  4  1  4]\n",
      " [12 10 41  6 15  2]\n",
      " [ 2  1  2  5  1  5]\n",
      " [ 5  1 15  1 11  2]\n",
      " [ 8  1  7  3  0 14]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAEmCAYAAADfpHMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlOD/Ae6bvxV1zB3sAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9eada25e80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.predict(n_gram_vectorizer, charn_model, test_data, my_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note**: The model performs really bad :_(, even worst than chance. Perhaps we can improve it by removing stopwords. 30% accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. TF-IDF\n",
    "Term Frequency - Inverse Document Frequency is useful to rank importance of words to documents"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tf_vect = TfidfVectorizer(\n",
    "    min_df=2, tokenizer=nltk.word_tokenize,\n",
    "    preprocessor=None, stop_words='english')\n",
    "train_data_features = tf_vect.fit_transform(train_data['plot'])\n",
    "\n",
    "tfidf_model = linear_model.LogisticRegression(n_jobs=1, C=1e5)\n",
    "tfidf_model = tfidf_model.fit(train_data_features, train_data['tag'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['caesar',\n",
       " 'cage',\n",
       " 'caine',\n",
       " 'cal',\n",
       " 'calhoun',\n",
       " 'california',\n",
       " 'californians',\n",
       " 'called',\n",
       " 'calling',\n",
       " 'callous']"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf_vect.get_feature_names()[1000:1010]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.465020576132\n",
      "confusion matrix\n",
      " [[23  2 12  2  2  1]\n",
      " [ 3 10 10  1  3  4]\n",
      " [ 9  6 49  0 21  1]\n",
      " [ 3  4  1  4  2  2]\n",
      " [ 1  2 20  0 11  1]\n",
      " [ 9  2  5  1  0 16]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAEmCAYAAADfpHMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9ead1708d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.predict(tf_vect, tfidf_model, test_data, my_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us do some analysis on the model now"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['comedy', 24.5992977289301],\n",
       " ['jewish', 21.111621307928754],\n",
       " ['mistaken', 20.906784380688975],\n",
       " ['kindergarten', 20.550050838076675],\n",
       " ['duo', 20.039310378863604],\n",
       " ['dimension', 19.730576933476108],\n",
       " ['beloved', 19.655031499523833],\n",
       " ['dealer', 19.190813922478519],\n",
       " ['street', 18.766362320888405],\n",
       " ['tramp', 18.567543420444927]]"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# words for the comedy genre\n",
    "comedy_tag_id = word_embeddings.get_tag_index(my_tags, \"comedy\")\n",
    "comedy_words = word_embeddings.most_influential_words(tfidf_model, tf_vect, comedy_tag_id, 3000)\n",
    "comedy_words[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['edward', 17.405228729634185],\n",
       " ['magical', 16.525441688448286],\n",
       " ['kingdom', 15.293991691029369],\n",
       " ['adventures', 14.936271113826516],\n",
       " ['percy', 14.809172140901389],\n",
       " ['moving', 14.203174472333552],\n",
       " ['land', 14.153599202700454],\n",
       " ['demon', 14.140054691071217],\n",
       " ['fan', 14.042094086775878],\n",
       " ['king', 13.98948584958028]]"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# words for the fantasy genre\n",
    "fantasy_tag_id = word_embeddings.get_tag_index(my_tags, \"fantasy\")\n",
    "fantasy_words = word_embeddings.most_influential_words(tfidf_model, tf_vect, fantasy_tag_id, 3000)\n",
    "fantasy_words[0:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let us observe what are the most influential words for a given movie plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['boy',\n",
       " 'loves',\n",
       " 'playing',\n",
       " 'come',\n",
       " 'worry',\n",
       " 'family',\n",
       " 'does',\n",
       " 'mother',\n",
       " 'action',\n",
       " 'quickly',\n",
       " 'tries',\n",
       " 'rid',\n",
       " 'ruthless',\n",
       " 'toy']"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_embeddings.most_influential_words_doc(train_data['plot'][0], fantasy_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note:** As you can see from the list above, there are words, such as *does* and *tries* that are not so relevant to the *fantasy* category. Therefore, we can still do better with this model. \n",
    "\n",
    "**Excercise:** Head to [scikit-learn](http://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html) and try to use different parameter for the TF-IDF vectorizer, such as modifying the *ngram_range*. It might improve the performance of the classifier. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Averaging Word Vectors (word2vec)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**SOURCE:** https://code.google.com/archive/p/word2vec/\n",
    "\n",
    "First we are going to use a pretrained model open source by Google (word2vec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:11:11,385 : INFO : loading projection weights from data/google_word2vec/GoogleNews-vectors-negative300.bin.gz\n",
      "2017-06-28 22:13:00,195 : INFO : loaded (3000000, 300) matrix from data/google_word2vec/GoogleNews-vectors-negative300.bin.gz\n",
      "2017-06-28 22:13:00,195 : INFO : precomputing L2-norms of word weight vectors\n"
     ]
    }
   ],
   "source": [
    "wv = gensim.models.KeyedVectors.load_word2vec_format(\n",
    "    \"data/google_word2vec/GoogleNews-vectors-negative300.bin.gz\",\n",
    "    binary=True)\n",
    "wv.init_sims(replace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise:** Create a function for exploring the word2vec file and find ways to visualize the results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Preprocessing:** Here we tokenize both the training and testing datasets before creating the vectors with word2vec. Think of this as a filtering step. In this case words with `len(word)<2` are removed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# test_tokenized = test_data.apply(lambda r: word2vec_helpers.w2v_tokenize_text(r['plot']), axis=1).values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train_tokenized = train_data.apply(lambda r: word2vec_helpers.w2v_tokenize_text(r['plot']), axis=1).values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# store the data as pickle for retrieving later\n",
    "#pickle_helpers.convert_to_pickle(train_tokenized, \"data/movie_reviews/train.p\")\n",
    "\n",
    "# retrieve pickled version of test and train tokenized\n",
    "test_tokenized = pickle_helpers.load_from_pickle(\"data/movie_reviews/test.p\")\n",
    "train_tokenized = pickle_helpers.load_from_pickle(\"data/movie_reviews/train.p\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ list(['Turkish', 'and', 'his', 'close', 'friend/accomplice', 'Tommy', 'get', 'pulled', 'into', 'the', 'world', 'of', 'match', 'fixing', 'by', 'the', 'notorious', 'Brick', 'Top', 'Things', 'get', 'complicated', 'when', 'the', 'boxer', 'they', 'had', 'lined', 'up', 'gets', 'badly', 'beaten', 'by', 'Pitt', \"'pikey\", 'slang', 'for', 'an', 'Irish', 'Gypsy', 'who', 'comes', 'into', 'the', 'equation', 'after', 'Turkish', 'an', 'unlicensed', 'boxing', 'promoter', 'wants', 'to', 'buy', 'caravan', 'off', 'the', 'Irish', 'Gypsies', 'They', 'then', 'try', 'to', 'convince', 'Pitt', 'not', 'only', 'to', 'fight', 'for', 'them', 'but', 'to', 'lose', 'for', 'them', 'too', 'Whilst', 'all', 'this', 'is', 'going', 'on', 'huge', 'diamond', 'heist', 'takes', 'place', 'and', 'fistful', 'of', 'motley', 'characters', 'enter', 'the', 'story', 'including', \"'Cousin\", 'Avi', \"'Boris\", 'The', 'Blade', \"'Franky\", 'Four', 'Fingers', 'and', \"'Bullet\", 'Tooth', 'Tony', 'Things', 'go', 'from', 'bad', 'to', 'worse', 'as', 'it', 'all', 'becomes', 'about', 'the', 'money', 'the', 'guns', 'and', 'the', 'damned', 'dog']),\n",
       "       list(['In', 'the', 'early', '1960', \"'s\", 'sixteen', 'year', 'old', 'Jenny', 'Mellor', 'lives', 'with', 'her', 'parents', 'in', 'the', 'London', 'suburb', 'of', 'Twickenham', 'On', 'her', 'father', \"'s\", 'wishes', 'everything', 'that', 'Jenny', 'does', 'is', 'in', 'the', 'sole', 'pursuit', 'of', 'being', 'accepted', 'into', 'Oxford', 'as', 'he', 'wants', 'her', 'to', 'have', 'better', 'life', 'than', 'he', 'Jenny', 'is', 'bright', 'pretty', 'hard', 'working', 'but', 'also', 'naturally', 'gifted', 'The', 'only', 'problems', 'her', 'father', 'may', 'perceive', 'in', 'her', 'life', 'is', 'her', 'issue', 'with', 'learning', 'Latin', 'and', 'her', 'dating', 'boy', 'named', 'Graham', 'who', 'is', 'nice', 'but', 'socially', 'awkward', 'Jenny', \"'s\", 'life', 'changes', 'after', 'she', 'meets', 'David', 'Goldman', 'man', 'over', 'twice', 'her', 'age', 'David', 'goes', 'out', 'of', 'his', 'way', 'to', 'show', 'Jenny', 'and', 'her', 'family', 'that', 'his', 'interest', 'in', 'her', 'is', 'not', 'improper', 'and', 'that', 'he', 'wants', 'solely', 'to', 'expose', 'her', 'to', 'cultural', 'activities', 'which', 'she', 'enjoys', 'Jenny', 'quickly', 'gets', 'accustomed', 'to', 'the', 'life', 'to', 'which', 'David', 'and', 'his', 'constant', 'companions', 'Danny', 'and', 'Helen', 'have', 'shown', 'her', 'and', 'Jenny', 'and', 'David', \"'s\", 'relationship', 'does', 'move', 'into', 'becoming', 'romantic', 'one', 'However', 'Jenny', 'slowly', 'learns', 'more', 'about', 'David', 'and', 'by', 'association', 'Danny', 'and', 'Helen', 'and', 'specifically', 'how', 'they', 'make', 'their', 'money', 'Jenny', 'has', 'to', 'decide', 'if', 'what', 'she', 'learns', 'about', 'them', 'and', 'leading', 'such', 'life', 'is', 'worth', 'forgoing', 'her', 'plans', 'of', 'higher', 'eduction', 'at', 'Oxford']),\n",
       "       list(['Ollie', 'Trinkie', 'is', 'publicist', 'who', 'has', 'great', 'girlfriend', 'Gertrude', 'whom', 'he', 'marries', 'and', 'they', 'are', 'expecting', 'baby', 'but', 'while', 'he', 'is', 'looking', 'forward', 'to', 'being', 'father', 'he', 'does', \"n't\", 'lighten', 'his', 'workload', 'Gertrude', 'gives', 'birth', 'but', 'dies', 'in', 'the', 'process', 'Ollie', 'does', \"n't\", 'live', 'up', 'to', 'his', 'responsibilities', 'as', 'father', 'Eventually', 'the', 'strain', 'and', 'pressure', 'of', 'losing', 'his', 'wife', 'and', 'being', 'father', 'gets', 'to', 'him', 'and', 'he', 'has', 'breakdown', 'which', 'leads', 'to', 'his', 'termination', 'So', 'with', 'nothing', 'much', 'to', 'do', 'he', 'tries', 'to', 'be', 'good', 'father', 'to', 'his', 'daughter', 'Gertie', 'He', 'also', 'meets', 'young', 'woman', 'name', 'Maya', 'who', 'likes', 'him', 'but', 'he', 'is', 'still', 'not', 'over', 'his', 'wife']),\n",
       "       list(['The', 'film', 'shows', 'the', 'day', 'when', 'Rui', 'and', 'Vani', 'first', 'met', 'It', 'was', 'at', 'their', 'wedding', 'with', 'other', 'partners', 'Vani', 'was', 'going', 'to', 'marry', 'Srgio', 'and', 'Rui', 'would', 'marry', 'Marta', 'in', 'the', 'same', 'church', 'following', 'Vani', \"'s\", 'marriage', 'While', 'waiting', 'for', 'the', 'ceremony', 'they', 'begin', 'to', 'talk', 'Complications', 'ensue']),\n",
       "       list(['Underneath', 'the', 'sands', 'of', 'Egypt', 'Anubis', 'an', 'ancient', 'evil', 'spirit', 'has', 'awakened', 'It', \"'s\", 'up', 'to', 'Yugi', 'who', 'defeated', 'Anubis', 'centuries', 'ago', 'to', 'use', 'his', 'skill', 'and', 'determination', 'to', 'rid', 'the', 'world', 'of', 'evil', 'once', 'again']),\n",
       "       list(['After', 'successful', 'mission', 'against', 'drug', 'lords', 'the', 'efficient', 'Captain', 'Damien', 'Tomaso', 'is', 'framed', 'at', 'home', 'with', 'three', 'kilograms', 'of', 'heroin', 'planted', 'by', 'the', 'police', 'in', 'his', 'kitchen', 'and', 'he', 'is', 'arrested', 'Meanwhile', 'group', 'of', 'teenagers', 'film', 'the', 'action', 'of', 'dirty', 'agents', 'led', 'by', 'Roland', 'from', 'the', 'security', 'agency', 'executing', 'policemen', 'in', 'their', 'car', 'and', 'then', 'leaving', 'the', 'car', 'with', 'the', 'corpses', 'in', 'the', '13th', 'District', 'to', 'blame', 'the', 'gangs', 'and', 'begin', 'civil', 'war', 'Behind', 'these', 'events', 'the', 'corrupt', 'chief', 'of', 'the', 'security', 'agency', 'Walter', 'Gassman', 'had', 'received', 'huge', 'amount', 'as', 'kickback', 'from', 'the', 'constructor', 'Harriburton', 'that', 'has', 'interest', 'to', 'construct', 'buildings', 'in', 'the', 'poor', 'area', 'and', 'uses', 'the', 'situation', 'to', 'force', 'the', 'President', 'of', 'France', 'to', 'authorize', 'to', 'nuke', 'five', 'towers', 'in', 'the', 'district', 'The', 'teenager', 'with', 'the', 'film', 'is', 'hunted', 'by', 'the', 'police', 'but', 'he', 'delivers', 'the', 'memory', 'card', 'to', 'Leito', 'meanwhile', 'Damien', 'calls', 'him', 'from', 'the', 'precinct', 'asking', 'for', 'help', 'The', 'friends', 'team-up', 'with', 'five', 'dangerous', 'bosses', 'to', 'gather', 'evidences', 'to', 'prove', 'to', 'the', 'president', 'that', 'Gassman', 'has', 'provoked', 'the', 'conflict', 'in', 'their', 'district']),\n",
       "       list(['Two', 'New', 'Yorkers', 'are', 'accused', 'of', 'murder', 'in', 'rural', 'Alabama', 'while', 'on', 'their', 'way', 'back', 'to', 'college', 'and', 'one', 'of', 'their', 'cousins', '--', 'an', 'inexperienced', 'loudmouth', 'lawyer', 'not', 'accustomed', 'to', 'Southern', 'rules', 'and', 'manners', '--', 'comes', 'in', 'to', 'defend', 'them']),\n",
       "       list(['man', 'finds', 'out', 'that', 'what', 'you', 'do', \"n't\", 'say', 'to', 'friend', 'is', 'just', 'as', 'important', 'as', 'what', 'you', 'do', 'is', 'this', 'story', 'of', 'how', 'far', 'you', 'can', 'bend', 'brotherly', 'bond', 'before', 'it', 'snaps', 'Since', 'college', 'confirmed', 'bachelor', 'Ronny', 'Vaughn', 'and', 'happily', 'married', 'Nick', 'James', 'have', 'been', 'through', 'thick', 'and', 'thin', 'Now', 'partners', 'in', 'an', 'auto', 'design', 'firm', 'the', 'two', 'pals', 'are', 'vying', 'to', 'land', 'dream', 'project', 'that', 'would', 'launch', 'their', 'company', 'Ronny', \"'s\", 'girlfriend', 'Beth', 'Connelly', 'and', 'Nick', \"'s\", 'wife', 'Geneva', 'Ryder', 'are', 'by', 'their', 'sides', 'But', 'Ronny', \"'s\", 'world', 'is', 'turned', 'upside', 'down', 'when', 'he', 'inadvertently', 'sees', 'Geneva', 'out', 'with', 'another', 'man', 'and', 'makes', 'it', 'his', 'mission', 'to', 'get', 'answers', 'As', 'the', 'amateur', 'investigation', 'dissolves', 'into', 'mayhem', 'he', 'learns', 'that', 'Nick', 'has', 'few', 'secrets', 'of', 'his', 'own', 'Now', 'with', 'the', 'clock', 'ticking', 'and', 'pressure', 'mounting', 'on', 'the', 'biggest', 'presentation', 'of', 'their', 'careers', 'Ronny', 'must', 'decide', 'what', 'will', 'happen', 'if', 'he', 'reveals', 'the', 'truth', 'to', 'his', 'best', 'friend']),\n",
       "       list(['Ben', 'Sobol', 'Psychiatrist', 'has', 'few', 'problems', 'His', 'son', 'spies', 'on', 'his', 'patients', 'when', 'they', 'open', 'up', 'their', 'heart', 'his', 'parents', 'do', \"n't\", 'want', 'to', 'attend', 'his', 'upcoming', 'wedding', 'and', 'his', 'patients', 'problems', 'do', \"n't\", 'challenge', 'him', 'at', 'all', 'Paul', 'Vitti', 'Godfather', 'has', 'few', 'problems', 'as', 'well', 'Sudden', 'anxiety', 'attacks', 'in', 'public', 'certain', 'disability', 'to', 'kill', 'people', 'and', 'his', 'best', 'part', 'ceasing', 'service', 'when', 'needed', 'One', 'day', 'Ben', 'unfortunately', 'crashes', 'into', 'one', 'of', 'Vitti', \"'s\", 'cars', 'The', 'exchange', 'of', 'Ben', \"'s\", 'business', 'card', 'is', 'followed', 'by', 'business', 'visit', 'of', 'Don', 'Paul', 'Vitti', 'himself', 'who', 'wants', 'to', 'be', 'free', 'of', 'inner', 'conflict', 'within', 'two', 'weeks', 'before', 'all', 'the', 'Mafia', 'Dons', 'meet', 'Now', 'Ben', 'Sobol', 'feels', 'somewhat', 'challenged', 'as', 'his', 'wedding', 'is', 'soon', 'his', 'only', 'patient', 'keeps', 'him', 'busy', 'by', 'regarding', 'Ben', \"'s\", 'duty', 'as', '24', 'hour', 'standby', 'and', 'the', 'feds', 'keep', 'forcing', 'him', 'to', 'spy', 'on', 'Paul', 'Vitti', 'And', 'how', 'do', 'you', 'treat', 'patient', 'who', 'usually', 'solves', 'problems', 'with', 'gun']),\n",
       "       list(['robot', 'malfunction', 'creates', 'havoc', 'and', 'terror', 'for', 'unsuspecting', 'vacationers', 'at', 'futuristic', 'adult-themed', 'amusement', 'park'])], dtype=object)"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_tokenized[0:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we have to convert the words into a distributed representation (in other words, each token must be conversed into vector). Since we are using word2vec the convesion can be done easily.\n",
    "\n",
    "Now that we have the vectors, we can obtain vectors for an entire document, by averaging word vectors. (Warning: very naive appraoch!)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![alt txt](https://docs.google.com/drawings/d/1m1Uod7QU4iSN7J3iqsTU6JRWdGTEJRZ2dA7rZrgNHgA/pub?w=942&h=375)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_word_average = word2vec_helpers.word_averaging_list(wv,train_tokenized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_word_average = word2vec_helpers.word_averaging_list(wv,test_tokenized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.05545449,  0.05910575,  0.04342501,  0.07156485, -0.06337085,\n",
       "       -0.04693267,  0.02318664, -0.0889559 ,  0.06699731,  0.08280275,\n",
       "       -0.02325549, -0.11232544, -0.01920838,  0.04011356, -0.11890385,\n",
       "        0.03739267,  0.0560286 ,  0.08226582,  0.02232919, -0.04910114,\n",
       "       -0.00497308,  0.04151743,  0.02096887, -0.00108779,  0.08315163,\n",
       "        0.00165053, -0.10190015,  0.0887218 ,  0.04294581,  0.02587495,\n",
       "       -0.01478133,  0.02559877, -0.03495957, -0.00126466,  0.03815735,\n",
       "        0.00277985, -0.00618378, -0.0127198 ,  0.04081891,  0.07088073,\n",
       "        0.09200442, -0.08722563,  0.11710441, -0.04684571, -0.02880416,\n",
       "       -0.00468458, -0.05068978, -0.00313703,  0.06019693,  0.06222438,\n",
       "       -0.01847018,  0.11511943, -0.00197043, -0.07153045, -0.03648838,\n",
       "        0.04271036, -0.02165449, -0.06775223,  0.01724528, -0.07145506,\n",
       "       -0.04500109,  0.0923021 , -0.07771284, -0.11304772, -0.03642419,\n",
       "       -0.02397631,  0.00947447,  0.06566761, -0.04847883,  0.07447033,\n",
       "        0.04596359,  0.00533706,  0.07440901,  0.01862758, -0.10930904,\n",
       "       -0.05191687,  0.09581147,  0.13321283,  0.02818674,  0.09916298,\n",
       "        0.00275499, -0.11262483,  0.03205725, -0.00655283, -0.03152488,\n",
       "       -0.09904213, -0.10709953,  0.10126126, -0.01183448,  0.00032518,\n",
       "        0.0346867 ,  0.05184319, -0.05309206, -0.13063876, -0.04088032,\n",
       "       -0.05249476,  0.04398097,  0.01841562, -0.02550774,  0.01531992,\n",
       "       -0.05255707, -0.05905483,  0.0204562 , -0.02640847, -0.05222479,\n",
       "       -0.03798182, -0.01838683, -0.08310767,  0.02429862, -0.06920657,\n",
       "        0.00807706,  0.02362138, -0.00099598, -0.03800139,  0.06231945,\n",
       "        0.03415101,  0.0561602 , -0.01016205,  0.08438386,  0.07533283,\n",
       "       -0.13695887, -0.01100177, -0.10242955,  0.03188829, -0.04064427,\n",
       "       -0.02914017, -0.06386539, -0.06155058, -0.00615242,  0.00615733,\n",
       "       -0.03677453, -0.10083614, -0.09915902, -0.01499352, -0.0428905 ,\n",
       "       -0.07342237,  0.00816098, -0.01920961, -0.05493748,  0.04522207,\n",
       "        0.07613138, -0.05877322,  0.0417468 , -0.01223148,  0.03953735,\n",
       "        0.01429599, -0.02857977, -0.07490996, -0.07850028,  0.03174346,\n",
       "        0.06443751,  0.01268353, -0.08018931,  0.06602806, -0.03417638,\n",
       "        0.00344341, -0.06512572, -0.08546869, -0.02787689, -0.0415597 ,\n",
       "       -0.02506224,  0.041864  ,  0.05230035,  0.01846709, -0.00877439,\n",
       "       -0.10832384,  0.03634784, -0.06550397,  0.03707062, -0.0529377 ,\n",
       "       -0.15230809, -0.06808924, -0.0105572 , -0.09720243, -0.02477943,\n",
       "       -0.0201604 ,  0.08111939, -0.06768905,  0.0002988 ,  0.0077303 ,\n",
       "       -0.04402705, -0.08054816, -0.00552738,  0.028412  , -0.03715033,\n",
       "       -0.04948618,  0.00337431,  0.03152485,  0.1285221 ,  0.04539691,\n",
       "        0.06230238,  0.00652147,  0.06113409,  0.02588119, -0.05525883,\n",
       "       -0.00190705, -0.04055991, -0.01651839, -0.08220793, -0.12156011,\n",
       "        0.05581678,  0.0460983 , -0.02329248,  0.02569922,  0.04048427,\n",
       "        0.01138046, -0.04218122,  0.01308791,  0.03205488, -0.0349777 ,\n",
       "       -0.01670968,  0.0374642 ,  0.01190601,  0.06392512, -0.10639601,\n",
       "        0.03272881,  0.10028993, -0.00271489, -0.10370444, -0.016493  ,\n",
       "       -0.01203196, -0.01359196, -0.0107542 , -0.03691358,  0.07650792,\n",
       "       -0.0165319 ,  0.05781083,  0.03541211,  0.06619678, -0.03761706,\n",
       "        0.04056923, -0.08521704,  0.04542974,  0.05765659,  0.01986002,\n",
       "       -0.0354447 ,  0.03333832, -0.05173995,  0.06272335,  0.01096975,\n",
       "        0.04950256, -0.00945826, -0.00653899, -0.1073802 ,  0.01505139,\n",
       "        0.00831405,  0.0658185 ,  0.06472088,  0.02112927, -0.0161819 ,\n",
       "        0.03234134,  0.01384802,  0.05193314,  0.05432371,  0.05399175,\n",
       "       -0.01561716,  0.07086278,  0.00388288, -0.04355494, -0.05447726,\n",
       "       -0.025471  ,  0.02247729, -0.08989162,  0.06539894,  0.08100092,\n",
       "        0.13270888, -0.06876201,  0.02076347, -0.05209624,  0.02469205,\n",
       "        0.06893449,  0.0873263 ,  0.07789882,  0.04711369,  0.0824732 ,\n",
       "       -0.08542187, -0.04682464, -0.15431304, -0.04475014, -0.01971488,\n",
       "        0.00911232, -0.0603571 , -0.0022078 ,  0.04911681,  0.04853116,\n",
       "        0.00946618, -0.12347805,  0.00823581,  0.00914641,  0.02190395,\n",
       "       -0.0688283 ,  0.03558169, -0.11885665,  0.02563123, -0.02871157,\n",
       "        0.0118563 ,  0.00246904, -0.02010039,  0.03884799, -0.0357489 ], dtype=float32)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_word_average[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Further exploration  of word2vec generated vectors\n",
    "Some tips here: https://radimrehurek.com/gensim/models/keyedvectors.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ -8.43075514e-02,   9.63514894e-02,   6.45992905e-02,\n",
       "         6.89789057e-02,   3.52193899e-02,   1.74272116e-02,\n",
       "        -6.05846494e-02,  -9.63514894e-02,  -7.00738132e-02,\n",
       "         2.60951947e-02,  -2.97448728e-02,   4.58034538e-02,\n",
       "        -4.89056818e-02,   3.55843566e-02,  -6.05846494e-02,\n",
       "        -3.67248803e-03,   2.93799043e-02,  -3.04748081e-02,\n",
       "        -3.10222600e-02,   4.96356152e-02,  -1.32118329e-01,\n",
       "        -6.60591647e-02,   1.71534847e-02,   6.67890981e-02,\n",
       "         7.81031027e-02,  -9.78113636e-02,  -7.25373439e-03,\n",
       "         4.31118160e-03,  -7.52746034e-03,  -9.48916152e-02,\n",
       "        -1.00001164e-01,  -3.50369066e-02,   8.50374922e-02,\n",
       "         1.72447264e-02,  -3.48544233e-02,  -5.21903895e-02,\n",
       "        -3.22996452e-02,   1.53286457e-02,  -4.58490755e-03,\n",
       "        -5.83948418e-02,   9.34317484e-02,   7.95629695e-02,\n",
       "        -2.14418564e-02,   2.40878724e-02,  -2.64601633e-02,\n",
       "        -8.72272924e-02,  -6.47247536e-04,  -6.60591647e-02,\n",
       "         9.63514894e-02,  -9.16981511e-03,  -4.41610999e-02,\n",
       "         2.59127114e-02,   1.31388396e-01,  -7.26285875e-02,\n",
       "         2.08031628e-02,   1.40512586e-02,  -1.79746617e-02,\n",
       "         7.00738132e-02,  -4.10588719e-02,  -1.13869943e-01,\n",
       "        -1.40512586e-02,  -2.64601633e-02,  -8.46725181e-02,\n",
       "        -3.04748081e-02,  -4.43435833e-02,   6.49642646e-02,\n",
       "         9.34317484e-02,  -2.68251300e-02,   5.72999381e-02,\n",
       "         4.85407114e-02,  -4.43435833e-02,  -2.82850023e-03,\n",
       "        -5.80298752e-02,  -4.12413590e-02,  -6.38693571e-02,\n",
       "        -1.71534847e-02,  -7.22636208e-02,   1.38687752e-02,\n",
       "         8.54024589e-02,  -5.62050343e-02,   5.62050343e-02,\n",
       "         1.49636781e-02,  -1.76781265e-03,  -5.98547123e-02,\n",
       "        -5.29203266e-02,   1.96170174e-02,   8.61323923e-02,\n",
       "        -3.01098414e-02,  -5.29203266e-02,   3.97814848e-02,\n",
       "         3.30752041e-03,  -1.35038078e-01,  -9.30667818e-02,\n",
       "         4.39101830e-04,  -1.21169299e-01,  -6.60591647e-02,\n",
       "         4.44804458e-03,   8.61323923e-02,   7.11687142e-03,\n",
       "         3.95990014e-02,   4.92706485e-02,  -5.36502600e-02,\n",
       "         5.14604561e-02,   6.02196828e-02,   1.00731105e-01,\n",
       "         2.39053890e-02,   2.66426466e-02,   1.29563557e-02,\n",
       "         2.07119212e-02,   1.46899521e-02,  -3.01098414e-02,\n",
       "        -3.10222600e-02,   8.94171000e-02,  -6.71540722e-02,\n",
       "         1.97082590e-02,  -1.54198883e-02,  -3.99639718e-02,\n",
       "        -5.43801971e-02,   4.41610999e-02,  -1.33578196e-01,\n",
       "        -8.71360581e-03,   1.19709425e-01,  -1.23359106e-01,\n",
       "         1.36862909e-02,  -3.32120657e-02,   8.04069627e-04,\n",
       "        -1.21169299e-01,   2.91974209e-02,   7.26285875e-02,\n",
       "         5.83948418e-02,  -5.25553562e-02,   8.72272924e-02,\n",
       "         3.30295824e-02,  -1.47446975e-01,  -2.17155814e-02,\n",
       "         9.53478273e-03,   4.01464552e-02,  -1.08121699e-02,\n",
       "        -2.29017269e-02,   9.23368409e-02,   1.05384439e-02,\n",
       "         6.70628250e-03,   4.72177053e-03,   7.11687133e-02,\n",
       "        -1.00731105e-01,  -2.60951947e-02,  -1.24887412e-03,\n",
       "         7.22636208e-02,   7.59132951e-02,   4.54384871e-02,\n",
       "        -3.55843566e-02,  -7.33585209e-02,  -1.08030461e-01,\n",
       "        -4.34311628e-02,   1.42337428e-02,  -1.44527242e-01,\n",
       "        -1.70622431e-02,   1.77921783e-02,   4.21537757e-02,\n",
       "         5.63419005e-03,  -7.55483285e-02,  -5.07305190e-02,\n",
       "         3.72267105e-02,  -3.06572914e-02,  -3.75916809e-02,\n",
       "        -4.81757447e-02,   8.97820666e-02,  -2.35404205e-02,\n",
       "         4.58034538e-02,   1.23176621e-02,  -1.00001164e-01,\n",
       "        -7.52746034e-03,  -5.91247790e-02,  -7.73731694e-02,\n",
       "         1.82483885e-02,   3.39420028e-02,   1.39417693e-01,\n",
       "        -7.08037466e-02,   6.79752463e-03,   6.02196828e-02,\n",
       "         1.04380779e-01,   3.64967771e-02,  -1.08121699e-02,\n",
       "         1.42337428e-02,   1.15421051e-02,  -6.58652745e-04,\n",
       "         8.10228437e-02,   3.13644159e-05,   1.45987107e-03,\n",
       "         9.48916189e-03,  -6.82489723e-02,   1.00001164e-01,\n",
       "        -1.34308144e-01,   2.33807470e-04,  -5.40152304e-02,\n",
       "        -5.07305190e-02,  -3.08397766e-02,  -3.86865847e-02,\n",
       "        -3.13872285e-02,  -4.78107780e-02,  -7.02562928e-03,\n",
       "        -3.44438315e-03,  -1.25001455e-02,  -5.58400676e-02,\n",
       "         6.16795532e-02,  -7.73731694e-02,  -2.13506147e-02,\n",
       "         3.95990014e-02,   2.57302281e-02,   1.98907424e-02,\n",
       "         2.97448728e-02,  -4.12413590e-02,  -6.60591647e-02,\n",
       "        -2.79200338e-02,   7.44534209e-02,   5.18254228e-02,\n",
       "        -8.61323923e-02,   1.14599876e-01,  -7.99279436e-02,\n",
       "        -2.88324542e-02,   2.19893083e-02,   3.01098414e-02,\n",
       "        -4.70808409e-02,   3.17521952e-02,  -7.00738132e-02,\n",
       "         6.93438724e-02,  -2.15330981e-02,  -6.45992905e-02,\n",
       "         1.61498226e-02,  -5.65700047e-02,  -2.21717916e-02,\n",
       "        -3.32120657e-02,  -8.07491131e-03,  -1.88870821e-02,\n",
       "         1.37775326e-02,  -2.70076152e-02,   1.02920912e-01,\n",
       "        -3.77741642e-02,   2.66882684e-03,   8.61323923e-02,\n",
       "        -7.11687142e-03,   3.72267105e-02,   6.43255701e-03,\n",
       "        -5.54751009e-02,  -4.19712923e-02,   1.24089038e-02,\n",
       "        -3.37595195e-02,   3.10222600e-02,  -2.21717916e-02,\n",
       "        -9.07857344e-03,   1.01461038e-01,   5.54751009e-02,\n",
       "        -7.52746034e-03,   5.32852933e-02,   5.76649085e-02,\n",
       "         3.19346786e-02,  -1.18979491e-01,  -4.43435833e-02,\n",
       "        -5.94897456e-02,  -3.86865847e-02,  -2.93799043e-02,\n",
       "         6.35043904e-02,  -1.25001455e-02,  -8.35776180e-02,\n",
       "         1.89783238e-02,   6.56941980e-02,  -3.79566476e-02,\n",
       "        -2.48178076e-02,   3.57668400e-02,   4.99549648e-03,\n",
       "         7.44534209e-02,  -1.54746339e-01,  -3.64967771e-02,\n",
       "         5.36502600e-02,   6.31394237e-02,  -3.10222600e-02,\n",
       "        -6.56941952e-03,  -2.07119212e-02,   5.10954857e-02,\n",
       "         2.90149376e-02,  -2.45212717e-03,  -6.89789057e-02,\n",
       "        -1.05840653e-01,   1.97082590e-02,   2.48178076e-02,\n",
       "        -1.03559606e-02,   3.95990014e-02,   3.24821323e-02,\n",
       "         1.66972745e-02,   1.44162271e-02,  -1.92520488e-02,\n",
       "        -3.21171619e-02,   4.05114219e-02,  -8.10228437e-02,\n",
       "        -8.68623257e-02,   5.94897456e-02,  -4.85407114e-02,\n",
       "         3.30295824e-02,  -3.04748081e-02,   6.45992905e-02], dtype=float32)"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wv.syn0norm[wv.vocab['Murtuza'].index] # (300,)\n",
    "#wv.vocab['woman'].index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('queen', 0.7118191719055176),\n",
       " ('monarch', 0.6189674139022827),\n",
       " ('princess', 0.5902431011199951),\n",
       " ('crown_prince', 0.5499460697174072),\n",
       " ('prince', 0.5377321243286133),\n",
       " ('kings', 0.5236844420433044),\n",
       " ('Queen_Consort', 0.5235945582389832),\n",
       " ('queens', 0.5181134343147278),\n",
       " ('sultan', 0.5098593235015869),\n",
       " ('monarchy', 0.5087411403656006)]"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wv.most_similar(positive=['woman', 'king'], negative=['man'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Donald_Trump', 0.492326557636261),\n",
       " ('Mexican', 0.43394654989242554),\n",
       " ('Aguilar_Zinser', 0.4289255142211914),\n",
       " ('México', 0.4275927245616913),\n",
       " ('Mexcio', 0.4245557487010956),\n",
       " ('Jorge_Castañeda', 0.42415398359298706),\n",
       " ('Celaya_Guanajuato', 0.4219090938568115),\n",
       " ('Guatemala', 0.41712844371795654),\n",
       " ('Bosques_de_las_Lomas', 0.41683387756347656),\n",
       " ('Peru', 0.4160282015800476)]"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wv.most_similar(positive=['GOP', 'Trump'], negative=['money'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For fun, let us try to manually compute the cosine similarity of some of our word vectors, which is computed as follows:\n",
    "![alt txt](https://docs.google.com/drawings/d/1QI_5_uxopOYgk2PzU_RMTUGKkoNhJp81aIz4SUwsHIM/pub?w=956&h=263)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.76924075586436913"
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_embeddings.cosine_measure(wv.syn0norm[wv.vocab['gun'].index], wv.syn0norm[wv.vocab['pistol'].index])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### KNN Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we train a KNN and a logistic regression classifier and observe how they perfom on these word-averaging document features\n",
    "\n",
    "Read about KNN- https://en.wikipedia.org/wiki/K-nearest_neighbors_algorithm\n",
    "\n",
    "![alt txt](https://docs.google.com/drawings/d/13xu1RP3pVem277_eT900af7iz_MvTQUrLoxzWAewb_o/pub?w=817&h=614)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='brute', leaf_size=30, metric='cosine',\n",
       "           metric_params=None, n_jobs=1, n_neighbors=3, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn_naive_dv = KNeighborsClassifier(n_neighbors=3, n_jobs=1, algorithm='brute', metric='cosine' )\n",
    "knn_naive_dv.fit(X_train_word_average, train_data.tag)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "predicted = knn_naive_dv.predict(X_test_word_average)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.38683127572\n",
      "confusion matrix\n",
      " [[22  4  6  1  6  3]\n",
      " [ 9  4  9  4  5  0]\n",
      " [18  4 44  3 13  4]\n",
      " [ 7  3  1  2  1  2]\n",
      " [13  0  9  1 11  1]\n",
      " [13  2  1  4  2 11]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAEmCAYAAADfpHMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKEEE0xBCKEEE0xBCKEEE0xBCKMH/A4YeQ5e7cv9PAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9db6e5a7b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.evaluate_prediction(predicted, test_data.tag, my_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "KNN doesn't perform well (only 38%), even lower than chance. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Scruffy but irresistibly attractive Yau Muk-yan, without a job or a place to live, moves in with sensitive, shy piano tuner Chan Kar-fu. Both are disturbed, then obsessed, by the amateurish piano playing of upstairs neighbour Mok Man-yee. Obsession turns to romance, and romance to fantasy. The film is structured in four \"movements\": two themes (Yau Muk-yan, Mok Man-yee), a duet (Yau Muk-yan & Mok Man-yee), and a set of variations (a wild fantasy of Chan Kar-fu in his new novel).'"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data.iloc()[56]['plot']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('just', 0.5038168430328369),\n",
       " ('but', 0.5023055076599121),\n",
       " ('in', 0.49800050258636475),\n",
       " ('the', 0.4957137107849121),\n",
       " ('so', 0.4889165759086609),\n",
       " ('actually', 0.4778650999069214),\n",
       " ('By_TBT_staff', 0.4743911623954773),\n",
       " ('even', 0.47195717692375183),\n",
       " ('really', 0.4685917794704437),\n",
       " ('anyway', 0.4660709500312805),\n",
       " ('vice_versa', 0.45745164155960083),\n",
       " ('one', 0.4536522626876831),\n",
       " ('sort', 0.4488521218299866),\n",
       " ('that', 0.4462577700614929),\n",
       " ('dreamy', 0.44418084621429443),\n",
       " ('You_EIG', 0.44372493028640747),\n",
       " ('kind', 0.442863404750824),\n",
       " ('then', 0.44108718633651733),\n",
       " ('obviously', 0.4360683262348175),\n",
       " ('Chan', 0.43542975187301636)]"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wv.most_similar(positive=[X_test_word_average[56]], restrict_vocab=100000, topn=30)[0:20]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The problem with the is that the average of this particular document fall under some area that is not related to the original document."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise:** Try to remove stopwords, the function is already provided. Try to figure out how to solve this using the current implementation of the function. Also, report the accuracy you achieved."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Doc2Vec\n",
    "Paper: https://cs.stanford.edu/~quocle/paragraph_vector.pdf\n",
    "It is a **semi-supervised** approach since a **weak label or tag** (you can put many) is introduced into training documents before modeling. Read more about semi-supervised appraoches here: https://en.wikipedia.org/wiki/Semi-supervised_learning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![alt txt](https://docs.google.com/drawings/d/1TT6jNhk8rVoWMQajb8AIXCtyug0t_YVVZrvFidDPUPU/pub?w=679&h=375)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "from gensim.models import Doc2Vec\n",
    "from gensim.models.doc2vec import TaggedDocument"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_tagged = train_data.apply(\n",
    "    lambda r: TaggedDocument(words=word_embeddings.tokenize_text(r['plot']), tags=[r.tag]), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_tagged = test_data.apply(\n",
    "    lambda r: TaggedDocument(words=word_embeddings.tokenize_text(r['plot']), tags=[r.tag]), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TaggedDocument(words=['troubled', 'psychologist', 'is', 'sent', 'to', 'investigate', 'the', 'crew', 'of', 'an', 'isolated', 'research', 'station', 'orbiting', 'bizarre', 'planet'], tags=['sci-fi'])"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_tagged.values[50]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainsent = train_tagged.values\n",
    "testsent = test_tagged.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ TaggedDocument(words=['turkish', 'and', 'his', 'close', 'friend/accomplice', 'tommy', 'get', 'pulled', 'into', 'the', 'world', 'of', 'match', 'fixing', 'by', 'the', 'notorious', 'brick', 'top', 'things', 'get', 'complicated', 'when', 'the', 'boxer', 'they', 'had', 'lined', 'up', 'gets', 'badly', 'beaten', 'by', 'pitt', \"'pikey\", 'slang', 'for', 'an', 'irish', 'gypsy', 'who', 'comes', 'into', 'the', 'equation', 'after', 'turkish', 'an', 'unlicensed', 'boxing', 'promoter', 'wants', 'to', 'buy', 'caravan', 'off', 'the', 'irish', 'gypsies', 'they', 'then', 'try', 'to', 'convince', 'pitt', 'not', 'only', 'to', 'fight', 'for', 'them', 'but', 'to', 'lose', 'for', 'them', 'too', 'whilst', 'all', 'this', 'is', 'going', 'on', 'huge', 'diamond', 'heist', 'takes', 'place', 'and', 'fistful', 'of', 'motley', 'characters', 'enter', 'the', 'story', 'including', \"'cousin\", 'avi', \"'boris\", 'the', 'blade', \"'franky\", 'four', 'fingers', 'and', \"'bullet\", 'tooth', 'tony', 'things', 'go', 'from', 'bad', 'to', 'worse', 'as', 'it', 'all', 'becomes', 'about', 'the', 'money', 'the', 'guns', 'and', 'the', 'damned', 'dog'], tags=['comedy']),\n",
       "       TaggedDocument(words=['in', 'the', 'early', '1960', \"'s\", 'sixteen', 'year', 'old', 'jenny', 'mellor', 'lives', 'with', 'her', 'parents', 'in', 'the', 'london', 'suburb', 'of', 'twickenham', 'on', 'her', 'father', \"'s\", 'wishes', 'everything', 'that', 'jenny', 'does', 'is', 'in', 'the', 'sole', 'pursuit', 'of', 'being', 'accepted', 'into', 'oxford', 'as', 'he', 'wants', 'her', 'to', 'have', 'better', 'life', 'than', 'he', 'jenny', 'is', 'bright', 'pretty', 'hard', 'working', 'but', 'also', 'naturally', 'gifted', 'the', 'only', 'problems', 'her', 'father', 'may', 'perceive', 'in', 'her', 'life', 'is', 'her', 'issue', 'with', 'learning', 'latin', 'and', 'her', 'dating', 'boy', 'named', 'graham', 'who', 'is', 'nice', 'but', 'socially', 'awkward', 'jenny', \"'s\", 'life', 'changes', 'after', 'she', 'meets', 'david', 'goldman', 'man', 'over', 'twice', 'her', 'age', 'david', 'goes', 'out', 'of', 'his', 'way', 'to', 'show', 'jenny', 'and', 'her', 'family', 'that', 'his', 'interest', 'in', 'her', 'is', 'not', 'improper', 'and', 'that', 'he', 'wants', 'solely', 'to', 'expose', 'her', 'to', 'cultural', 'activities', 'which', 'she', 'enjoys', 'jenny', 'quickly', 'gets', 'accustomed', 'to', 'the', 'life', 'to', 'which', 'david', 'and', 'his', 'constant', 'companions', 'danny', 'and', 'helen', 'have', 'shown', 'her', 'and', 'jenny', 'and', 'david', \"'s\", 'relationship', 'does', 'move', 'into', 'becoming', 'romantic', 'one', 'however', 'jenny', 'slowly', 'learns', 'more', 'about', 'david', 'and', 'by', 'association', 'danny', 'and', 'helen', 'and', 'specifically', 'how', 'they', 'make', 'their', 'money', 'jenny', 'has', 'to', 'decide', 'if', 'what', 'she', 'learns', 'about', 'them', 'and', 'leading', 'such', 'life', 'is', 'worth', 'forgoing', 'her', 'plans', 'of', 'higher', 'eduction', 'at', 'oxford'], tags=['romance']),\n",
       "       TaggedDocument(words=['ollie', 'trinkie', 'is', 'publicist', 'who', 'has', 'great', 'girlfriend', 'gertrude', 'whom', 'he', 'marries', 'and', 'they', 'are', 'expecting', 'baby', 'but', 'while', 'he', 'is', 'looking', 'forward', 'to', 'being', 'father', 'he', 'does', \"n't\", 'lighten', 'his', 'workload', 'gertrude', 'gives', 'birth', 'but', 'dies', 'in', 'the', 'process', 'ollie', 'does', \"n't\", 'live', 'up', 'to', 'his', 'responsibilities', 'as', 'father', 'eventually', 'the', 'strain', 'and', 'pressure', 'of', 'losing', 'his', 'wife', 'and', 'being', 'father', 'gets', 'to', 'him', 'and', 'he', 'has', 'breakdown', 'which', 'leads', 'to', 'his', 'termination', 'so', 'with', 'nothing', 'much', 'to', 'do', 'he', 'tries', 'to', 'be', 'good', 'father', 'to', 'his', 'daughter', 'gertie', 'he', 'also', 'meets', 'young', 'woman', 'name', 'maya', 'who', 'likes', 'him', 'but', 'he', 'is', 'still', 'not', 'over', 'his', 'wife'], tags=['comedy']),\n",
       "       ...,\n",
       "       TaggedDocument(words=['the', 'factual', 'story', 'of', 'spaniard', 'ramon', 'sampedro', 'who', 'fought', 'thirty-year', 'campaign', 'in', 'favor', 'of', 'euthanasia', 'and', 'his', 'own', 'right', 'to', 'die'], tags=['romance']),\n",
       "       TaggedDocument(words=['in', 'south', 'boston', 'the', 'state', 'police', 'force', 'is', 'waging', 'war', 'on', 'irish-american', 'organized', 'crime', 'young', 'undercover', 'cop', 'billy', 'costigan', 'leonardo', 'dicaprio', 'is', 'assigned', 'to', 'infiltrate', 'the', 'mob', 'syndicate', 'run', 'by', 'gangland', 'chief', 'frank', 'costello', 'jack', 'nicholson', 'while', 'billy', 'quickly', 'gains', 'costello', \"'s\", 'confidence', 'colin', 'sullivan', 'matt', 'damon', 'hardened', 'young', 'criminal', 'who', 'has', 'infiltrated', 'the', 'state', 'police', 'as', 'an', 'informer', 'for', 'the', 'syndicate', 'is', 'rising', 'to', 'position', 'of', 'power', 'in', 'the', 'special', 'investigation', 'unit', 'each', 'man', 'becomes', 'deeply', 'consumed', 'by', 'their', 'double', 'lives', 'gathering', 'information', 'about', 'the', 'plans', 'and', 'counter-plans', 'of', 'the', 'operations', 'they', 'have', 'penetrated', 'but', 'when', 'it', 'becomes', 'clear', 'to', 'both', 'the', 'mob', 'and', 'the', 'police', 'that', 'there', 'is', 'mole', 'in', 'their', 'midst', 'billy', 'and', 'colin', 'are', 'suddenly', 'in', 'danger', 'of', 'being', 'caught', 'and', 'exposed', 'to', 'the', 'enemy', 'and', 'each', 'must', 'race', 'to', 'uncover', 'the', 'identity', 'of', 'the', 'other', 'man', 'in', 'time', 'to', 'save', 'themselves', 'but', 'is', 'either', 'willing', 'to', 'turn', 'on', 'their', 'friends', 'and', 'comrades', 'they', \"'ve\", 'made', 'during', 'their', 'long', 'stints', 'undercover'], tags=['action']),\n",
       "       TaggedDocument(words=['the', 'topalovic', 'family', 'has', 'been', 'in', 'the', 'burial', 'business', 'for', 'generations', 'when', 'the', 'old', '150', 'yrs', 'old', 'pantelija', 'dies', 'five', 'generations', 'of', 'his', 'heirs', 'start', 'to', 'fight', 'for', 'the', 'inheritance'], tags=['comedy'])], dtype=object)"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainsent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:28,285 : WARNING : consider setting layer size to a multiple of 4 for greater performance\n",
      "2017-06-28 22:13:28,286 : INFO : collecting all words and their counts\n",
      "2017-06-28 22:13:28,287 : INFO : PROGRESS: at example #0, processed 0 words (0/s), 0 word types, 0 tags\n",
      "2017-06-28 22:13:28,324 : INFO : collected 17168 word types and 6 unique tags from a corpus of 2184 examples and 150640 words\n",
      "2017-06-28 22:13:28,325 : INFO : Loading a fresh vocabulary\n",
      "2017-06-28 22:13:28,335 : INFO : min_count=5 retains 3631 unique words (21% of original 17168, drops 13537)\n",
      "2017-06-28 22:13:28,336 : INFO : min_count=5 leaves 128953 word corpus (85% of original 150640, drops 21687)\n",
      "2017-06-28 22:13:28,345 : INFO : deleting the raw counts dictionary of 17168 items\n",
      "2017-06-28 22:13:28,346 : INFO : sample=0.001 downsamples 43 most-common words\n",
      "2017-06-28 22:13:28,347 : INFO : downsampling leaves estimated 92782 word corpus (72.0% of prior 128953)\n",
      "2017-06-28 22:13:28,348 : INFO : estimated required memory for 3631 words and 5 dimensions: 1962060 bytes\n",
      "2017-06-28 22:13:28,357 : INFO : resetting layer weights\n",
      "2017-06-28 22:13:28,407 : INFO : training model with 1 workers on 3631 vocabulary and 5 features, using sg=0 hs=0 sample=0.001 negative=5 window=5\n",
      "2017-06-28 22:13:29,414 : INFO : PROGRESS: at 42.23% examples, 799030 words/s, in_qsize 2, out_qsize 0\n",
      "2017-06-28 22:13:30,419 : INFO : PROGRESS: at 84.13% examples, 795161 words/s, in_qsize 2, out_qsize 0\n",
      "2017-06-28 22:13:30,797 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-06-28 22:13:30,798 : INFO : training on 3012800 raw words (1899475 effective words) took 2.4s, 795208 effective words/s\n"
     ]
    }
   ],
   "source": [
    "# train the model with simple gensim doc2vec api\n",
    "# size (dimensions of features)\n",
    "# dm is the training algorithm (distribute memory (1) or distributed bag of words)\n",
    "doc2vec_model = Doc2Vec(trainsent, workers=1, size=5, iter=20, dm=1)\n",
    "\n",
    "train_targets, train_regressors = zip(\n",
    "    *[(doc.tags[0], doc2vec_model.infer_vector(doc.words, steps=20)) for doc in trainsent])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_targets, test_regressors = zip(\n",
    "    *[(doc.tags[0], doc2vec_model.infer_vector(doc.words, steps=20)) for doc in testsent])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us see what our data looks like after training and transormations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('comedy', 'romance', 'sci-fi', 'comedy', 'fantasy')"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_targets[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([-0.30981189,  0.78878409,  0.27795053, -1.40723765, -0.78624779], dtype=float32),\n",
       " array([-0.14812587,  0.67659032,  0.30580592, -0.62821662, -1.07823062], dtype=float32),\n",
       " array([-0.40741968,  0.41884324,  0.28939512, -0.16581796, -0.11567296], dtype=float32),\n",
       " array([-0.16268265,  0.93405986,  0.48112747, -0.88357091, -0.70414579], dtype=float32),\n",
       " array([ 0.06507298,  0.17371821,  0.09560882, -0.10981997, -0.2437499 ], dtype=float32))"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_regressors[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "d2v_model = linear_model.LogisticRegression(n_jobs=1, C=1e5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "d2v_model = d2v_model.fit(train_regressors, train_targets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.390946502058\n",
      "confusion matrix\n",
      " [[15  0 12  1  2 12]\n",
      " [ 5  0 14  3  5  4]\n",
      " [10  1 58  0 14  3]\n",
      " [ 2  0  6  1  3  4]\n",
      " [ 2  0 22  0  9  2]\n",
      " [ 8  0 12  1  0 12]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAEmCAYAAADfpHMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9db6d393c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.evaluate_prediction(d2v_model.predict(test_regressors), test_targets, my_tags,title=str(doc2vec_model))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let us train KNN model with doc2vec. You will observe how poor the model performs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:33,504 : INFO : precomputing L2-norms of doc weight vectors\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.201646090535\n",
      "confusion matrix\n",
      " [[ 0 37  1  0  3  1]\n",
      " [ 3 26  0  0  2  0]\n",
      " [ 2 56 14  1 12  1]\n",
      " [ 0 15  0  0  1  0]\n",
      " [ 0 24  4  0  6  1]\n",
      " [ 1 28  0  0  1  3]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAEmCAYAAADfpHMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDoQyDYIgyIFQpkEQBDnw/78sSqNVR9zpAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d942be240>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "knn_test_predictions = [\n",
    "    doc2vec_model.docvecs.most_similar([pred_vec], topn=1)[0][0]\n",
    "    for pred_vec in test_regressors\n",
    "]\n",
    "word_embeddings.evaluate_prediction(knn_test_predictions, test_targets,my_tags, str(doc2vec_model))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since doc2vec also gives us a vector for each genre, we can also observe which genres are similar to each other"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('sci-fi', 0.955485463142395),\n",
       " ('animation', 0.9373874068260193),\n",
       " ('fantasy', 0.5044772624969482),\n",
       " ('comedy', 0.35625192523002625),\n",
       " ('romance', 0.09751249849796295)]"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doc2vec_model.docvecs.most_similar('action')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('animation', 0.6903013586997986),\n",
       " ('romance', 0.6896251440048218),\n",
       " ('action', 0.5044772624969482),\n",
       " ('sci-fi', 0.47834616899490356),\n",
       " ('comedy', -0.18462462723255157)]"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doc2vec_model.docvecs.most_similar('fantasy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since words and categories fall in the same vector space, it is possible to also observe which words sorround a tag. Notice how well those words descrive the tag, sci-fi in this case"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('years', 0.9583119750022888),\n",
       " ('fell', 0.9185272455215454),\n",
       " ('regina', 0.9117251634597778),\n",
       " ('marks', 0.9067319631576538),\n",
       " ('infected', 0.9006549119949341),\n",
       " ('cunning', 0.8974300622940063),\n",
       " ('vampires', 0.892426609992981),\n",
       " ('skywalker', 0.8912185430526733),\n",
       " ('cyborg', 0.8903250098228455),\n",
       " ('unfolds', 0.8898645043373108)]"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doc2vec_model.most_similar([doc2vec_model.docvecs['fantasy']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "pylab.rcParams['figure.figsize'] = (15, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "#doc2vec_model.wv.vocab\n",
    "doc2vec_words = word2vec_helpers.unpack_words_from_doc_vector(doc2vec_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3MAAAJCCAYAAABqGJ7GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nH0PkOK0i3Lm+AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d942d4780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word2vec_helpers.visualize_vectors(doc2vec_model, doc2vec_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise:** Try to visualize words and find interesting clusters. There are so many creative ways to visualize these powerful vector representations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Deep IR"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Paper: https://arxiv.org/pdf/1504.07295v3.pdf\n",
    "\n",
    "This approach is very simple: Train a model for each class or tag and then see which model fits each plot the best using a technique called the [Bayes' Theorem](https://en.wikipedia.org/wiki/Bayes%27_theorem). \n",
    "\n",
    "Pre-processing: We just clean non-alphanumeric characters and split reviews by sentences."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The corpus is small so can be read into memory\n",
    "# here we split plot into sentences\n",
    "revtrain = list(deepir.plots(\"training\", train_data, test_data))\n",
    "revtest = list(deepir.plots(\"test\",train_data, test_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'x': [['turkish',\n",
       "    'and',\n",
       "    'his',\n",
       "    'close',\n",
       "    'friend',\n",
       "    'accomplice',\n",
       "    'tommy',\n",
       "    'get',\n",
       "    'pulled',\n",
       "    'into',\n",
       "    'the',\n",
       "    'world',\n",
       "    'of',\n",
       "    'match',\n",
       "    'fixing',\n",
       "    'by',\n",
       "    'the',\n",
       "    'notorious',\n",
       "    'brick',\n",
       "    'top'],\n",
       "   ['things',\n",
       "    'get',\n",
       "    'complicated',\n",
       "    'when',\n",
       "    'the',\n",
       "    'boxer',\n",
       "    'they',\n",
       "    'had',\n",
       "    'lined',\n",
       "    'up',\n",
       "    'gets',\n",
       "    'badly',\n",
       "    'beaten',\n",
       "    'by',\n",
       "    'pitt',\n",
       "    'pikey',\n",
       "    'slang',\n",
       "    'for',\n",
       "    'an',\n",
       "    'irish',\n",
       "    'gypsy',\n",
       "    'who',\n",
       "    'comes',\n",
       "    'into',\n",
       "    'the',\n",
       "    'equation',\n",
       "    'after',\n",
       "    'turkish',\n",
       "    'an',\n",
       "    'unlicensed',\n",
       "    'boxing',\n",
       "    'promoter',\n",
       "    'wants',\n",
       "    'to',\n",
       "    'buy',\n",
       "    'caravan',\n",
       "    'off',\n",
       "    'the',\n",
       "    'irish',\n",
       "    'gypsies'],\n",
       "   ['they',\n",
       "    'then',\n",
       "    'try',\n",
       "    'to',\n",
       "    'convince',\n",
       "    'pitt',\n",
       "    'not',\n",
       "    'only',\n",
       "    'to',\n",
       "    'fight',\n",
       "    'for',\n",
       "    'them',\n",
       "    'but',\n",
       "    'to',\n",
       "    'lose',\n",
       "    'for',\n",
       "    'them',\n",
       "    'too'],\n",
       "   ['whilst',\n",
       "    'all',\n",
       "    'this',\n",
       "    'is',\n",
       "    'going',\n",
       "    'on',\n",
       "    'huge',\n",
       "    'diamond',\n",
       "    'heist',\n",
       "    'takes',\n",
       "    'place',\n",
       "    'and',\n",
       "    'fistful',\n",
       "    'of',\n",
       "    'motley',\n",
       "    'characters',\n",
       "    'enter',\n",
       "    'the',\n",
       "    'story',\n",
       "    'including',\n",
       "    'cousin',\n",
       "    'avi',\n",
       "    'boris',\n",
       "    'the',\n",
       "    'blade',\n",
       "    'franky',\n",
       "    'four',\n",
       "    'fingers',\n",
       "    'and',\n",
       "    'bullet',\n",
       "    'tooth',\n",
       "    'tony'],\n",
       "   ['things',\n",
       "    'go',\n",
       "    'from',\n",
       "    'bad',\n",
       "    'to',\n",
       "    'worse',\n",
       "    'as',\n",
       "    'it',\n",
       "    'all',\n",
       "    'becomes',\n",
       "    'about',\n",
       "    'the',\n",
       "    'money',\n",
       "    'the',\n",
       "    'guns',\n",
       "    'and',\n",
       "    'the',\n",
       "    'damned',\n",
       "    'dog']],\n",
       "  'y': 'comedy'}]"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "revtrain[0:1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "# shuffle training set for unbiased word2vec training\n",
    "np.random.shuffle(revtrain)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us look at an example from the category sci-fi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['in',\n",
       " 'fascist',\n",
       " 'future',\n",
       " 'where',\n",
       " 'all',\n",
       " 'forms',\n",
       " 'of',\n",
       " 'feeling',\n",
       " 'are',\n",
       " 'illegal',\n",
       " 'man',\n",
       " 'in',\n",
       " 'charge',\n",
       " 'of',\n",
       " 'enforcing',\n",
       " 'the',\n",
       " 'law',\n",
       " 'rises',\n",
       " 'to',\n",
       " 'overthrow',\n",
       " 'the',\n",
       " 'system']"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "next(deepir.tag_sentences(revtrain, [\"sci-fi\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the following steps, we train 6 word2vec models from scratch instead of using the Google pretrained word2vec as in the previous excercises"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "## training\n",
    "from gensim.models import Word2Vec\n",
    "import multiprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "## 1. create a w2v learner \n",
    "basemodel = Word2Vec(\n",
    "    workers=multiprocessing.cpu_count(), # use your cores\n",
    "    iter=100, # iter = sweeps of SGD through the data; more is better\n",
    "    hs=1, negative=0, # we only have scoring for the hierarchical softmax setup\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Word2Vec(vocab=0, size=100, alpha=0.025)\n"
     ]
    }
   ],
   "source": [
    "print(basemodel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "100"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "basemodel.iter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:38,994 : INFO : collecting all words and their counts\n",
      "2017-06-28 22:13:38,996 : INFO : PROGRESS: at sentence #0, processed 0 words, keeping 0 word types\n",
      "2017-06-28 22:13:39,056 : INFO : collected 17544 word types from a corpus of 148422 raw words and 7880 sentences\n",
      "2017-06-28 22:13:39,057 : INFO : Loading a fresh vocabulary\n",
      "2017-06-28 22:13:39,070 : INFO : min_count=5 retains 3647 unique words (20% of original 17544, drops 13897)\n",
      "2017-06-28 22:13:39,070 : INFO : min_count=5 leaves 126255 word corpus (85% of original 148422, drops 22167)\n",
      "2017-06-28 22:13:39,080 : INFO : deleting the raw counts dictionary of 17544 items\n",
      "2017-06-28 22:13:39,081 : INFO : sample=0.001 downsamples 42 most-common words\n",
      "2017-06-28 22:13:39,081 : INFO : downsampling leaves estimated 91197 word corpus (72.2% of prior 126255)\n",
      "2017-06-28 22:13:39,082 : INFO : estimated required memory for 3647 words and 100 dimensions: 5470500 bytes\n",
      "2017-06-28 22:13:39,085 : INFO : constructing a huffman tree from 3647 words\n",
      "2017-06-28 22:13:39,155 : INFO : built huffman tree with maximum node depth 15\n",
      "2017-06-28 22:13:39,157 : INFO : resetting layer weights\n"
     ]
    }
   ],
   "source": [
    "# 2. build a vocabulary\n",
    "basemodel.build_vocab(deepir.tag_sentences(revtrain, my_tags)) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "from copy import deepcopy\n",
    "genremodels = [deepcopy(basemodel) for i in range(len(my_tags))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:39,592 : INFO : training model with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 negative=0 window=5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sci-fi genre ( 1157 )\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:40,340 : INFO : worker thread finished; awaiting finish of 3 more threads\n",
      "2017-06-28 22:13:40,342 : INFO : worker thread finished; awaiting finish of 2 more threads\n",
      "2017-06-28 22:13:40,345 : INFO : worker thread finished; awaiting finish of 1 more threads\n",
      "2017-06-28 22:13:40,345 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-06-28 22:13:40,346 : INFO : training on 2163100 raw words (1331086 effective words) took 0.7s, 1780320 effective words/s\n",
      "2017-06-28 22:13:40,347 : INFO : training model with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 negative=0 window=5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "action genre ( 1370 )\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:41,269 : INFO : worker thread finished; awaiting finish of 3 more threads\n",
      "2017-06-28 22:13:41,272 : INFO : worker thread finished; awaiting finish of 2 more threads\n",
      "2017-06-28 22:13:41,276 : INFO : worker thread finished; awaiting finish of 1 more threads\n",
      "2017-06-28 22:13:41,278 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-06-28 22:13:41,278 : INFO : training on 2577000 raw words (1598414 effective words) took 0.9s, 1735829 effective words/s\n",
      "2017-06-28 22:13:41,280 : INFO : training model with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 negative=0 window=5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "comedy genre ( 2375 )\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:42,287 : INFO : PROGRESS: at 62.38% examples, 1723566 words/s, in_qsize 7, out_qsize 0\n",
      "2017-06-28 22:13:42,862 : INFO : worker thread finished; awaiting finish of 3 more threads\n",
      "2017-06-28 22:13:42,863 : INFO : worker thread finished; awaiting finish of 2 more threads\n",
      "2017-06-28 22:13:42,866 : INFO : worker thread finished; awaiting finish of 1 more threads\n",
      "2017-06-28 22:13:42,871 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-06-28 22:13:42,872 : INFO : training on 4515700 raw words (2765161 effective words) took 1.6s, 1744093 effective words/s\n",
      "2017-06-28 22:13:42,873 : INFO : training model with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 negative=0 window=5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "fantasy genre ( 707 )\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:43,342 : INFO : worker thread finished; awaiting finish of 3 more threads\n",
      "2017-06-28 22:13:43,345 : INFO : worker thread finished; awaiting finish of 2 more threads\n",
      "2017-06-28 22:13:43,347 : INFO : worker thread finished; awaiting finish of 1 more threads\n",
      "2017-06-28 22:13:43,349 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-06-28 22:13:43,349 : INFO : training on 1340700 raw words (822429 effective words) took 0.5s, 1748793 effective words/s\n",
      "2017-06-28 22:13:43,351 : INFO : training model with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 negative=0 window=5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "animation genre ( 867 )\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:43,952 : INFO : worker thread finished; awaiting finish of 3 more threads\n",
      "2017-06-28 22:13:43,958 : INFO : worker thread finished; awaiting finish of 2 more threads\n",
      "2017-06-28 22:13:43,962 : INFO : worker thread finished; awaiting finish of 1 more threads\n",
      "2017-06-28 22:13:43,963 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-06-28 22:13:43,963 : INFO : training on 1604300 raw words (972786 effective words) took 0.6s, 1603441 effective words/s\n",
      "2017-06-28 22:13:43,966 : INFO : training model with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 negative=0 window=5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "romance genre ( 1404 )\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:44,910 : INFO : worker thread finished; awaiting finish of 3 more threads\n",
      "2017-06-28 22:13:44,914 : INFO : worker thread finished; awaiting finish of 2 more threads\n",
      "2017-06-28 22:13:44,915 : INFO : worker thread finished; awaiting finish of 1 more threads\n",
      "2017-06-28 22:13:44,917 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-06-28 22:13:44,917 : INFO : training on 2641400 raw words (1630692 effective words) took 0.9s, 1725774 effective words/s\n"
     ]
    }
   ],
   "source": [
    "# 3. training for each model or each tag\n",
    "for i in range(len(my_tags)):\n",
    "    slist = list(deepir.tag_sentences(revtrain, my_tags[i]))\n",
    "    print(my_tags[i], \"genre (\", len(slist), \")\")\n",
    "    genremodels[i].train(  slist, total_examples=len(slist), epochs=basemodel.iter)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we will compute the most likely class for each plot using Bayes' Theorem  formula"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![alt text](https://docs.google.com/drawings/d/1SRnoMZ6TzZEfswazFhvxcQTrm2EC9lBn3qBVxPpRfXs/pub?w=866&h=424)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-06-28 22:13:44,932 : INFO : scoring sentences with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 and negative=0\n",
      "2017-06-28 22:13:44,934 : INFO : reached end of input; waiting to finish 9 outstanding jobs\n",
      "2017-06-28 22:13:44,945 : INFO : scoring 894 sentences took 0.0s, 73422 sentences/s\n",
      "2017-06-28 22:13:44,946 : INFO : scoring sentences with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 and negative=0\n",
      "2017-06-28 22:13:44,949 : INFO : reached end of input; waiting to finish 9 outstanding jobs\n",
      "2017-06-28 22:13:44,960 : INFO : scoring 894 sentences took 0.0s, 66932 sentences/s\n",
      "2017-06-28 22:13:44,960 : INFO : scoring sentences with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 and negative=0\n",
      "2017-06-28 22:13:44,961 : INFO : reached end of input; waiting to finish 9 outstanding jobs\n",
      "2017-06-28 22:13:44,971 : INFO : scoring 894 sentences took 0.0s, 83436 sentences/s\n",
      "2017-06-28 22:13:44,972 : INFO : scoring sentences with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 and negative=0\n",
      "2017-06-28 22:13:44,974 : INFO : reached end of input; waiting to finish 9 outstanding jobs\n",
      "2017-06-28 22:13:44,985 : INFO : scoring 894 sentences took 0.0s, 74658 sentences/s\n",
      "2017-06-28 22:13:44,985 : INFO : scoring sentences with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 and negative=0\n",
      "2017-06-28 22:13:44,986 : INFO : reached end of input; waiting to finish 9 outstanding jobs\n",
      "2017-06-28 22:13:44,996 : INFO : scoring 894 sentences took 0.0s, 83549 sentences/s\n",
      "2017-06-28 22:13:44,997 : INFO : scoring sentences with 4 workers on 3647 vocabulary and 100 features, using sg=0 hs=1 sample=0.001 and negative=0\n",
      "2017-06-28 22:13:44,998 : INFO : reached end of input; waiting to finish 9 outstanding jobs\n",
      "2017-06-28 22:13:45,010 : INFO : scoring 894 sentences took 0.0s, 72172 sentences/s\n"
     ]
    }
   ],
   "source": [
    "probs = deepir.docprob( [r['x'] for r in revtest], genremodels )  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>doc</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.140609</td>\n",
       "      <td>0.136355</td>\n",
       "      <td>0.314262</td>\n",
       "      <td>0.057756</td>\n",
       "      <td>0.284332</td>\n",
       "      <td>0.066686</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.025672</td>\n",
       "      <td>0.333158</td>\n",
       "      <td>0.311374</td>\n",
       "      <td>0.018236</td>\n",
       "      <td>0.000245</td>\n",
       "      <td>0.311316</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            0         1         2         3         4         5\n",
       "doc                                                            \n",
       "0    0.140609  0.136355  0.314262  0.057756  0.284332  0.066686\n",
       "1    0.025672  0.333158  0.311374  0.018236  0.000245  0.311316"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "probs[0:2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = probs.idxmax(axis=1).apply(lambda x: my_tags[x])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "doc\n",
       "0    comedy\n",
       "1    action\n",
       "dtype: object"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions[0:2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f9d8e64ca58>"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtwAAAFYCAYAAABgYCs6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XucXGV9+PHP14SbQEEFU+5BRbtxLWijaBvrbmMRRIu2\nKC6goqs0/Ulsa2tBVwuoq1BqveAlRZeCgos3VAQKWMwqEVFAuQRWlHILYMULIgEEEr6/P86zMFn2\nMkl2ZrIzn/frta/MnPOcc77nPDOT7zzznOeJzESSJElSYzyh1QFIkiRJ7cyEW5IkSWogE25JkiSp\ngUy4JUmSpAYy4ZYkSZIayIRbkiRJaiATbkltKSJOi4gPtDqOVpvqOkTEERGxotkxSVKnMeGW1FAR\ncUtEPBARqyPi7og4LyJ2a3VctSIiI+IZrY5jNouIkYh4S6vjkKRNkQm3pGZ4ZWZuA+wE/AI4ucXx\nNExU/GwdJyLmtjqGDTWbY5e0afA/BUlNk5m/B74CLBhbFhHbRcTnIuKXEXFrRLxnLGGNiE9HxFdr\nyp4YEReXpLYnIm6PiHdHxK9KS/phkx07It4aETdGxG8i4pyI2Lks/24pcnVphT9kgm3nRMSHy3Fu\njoijSqv43LJ+JCIGI+J7wP3A0yJi53Kc35TjvrVmf+t08xg7l5rnt0TEuyLi+vKrwH9FxJY1618R\nEVdFxG8j4tKI+OOadc+NiB9FxL0R8UXg0e0mvzTxiYi4JyJ+EhGLy8LXRMSV4wq+IyK+McEOBoEX\nA58o1/ATZXlGxNsi4mfAzyJifu11q7l2b6l5/uaIGC3nfWFE7DFF4G8or5lfR8R7y3V7aVn3hIg4\nJiL+t6z/UkQ8uawbi+ONEXFbqdeBmv0eFxFfiYgzIuJ3wBHT7G/LUvbXpU4uj4h501x3SR3EhFtS\n00TEE4FDgMtqFp8MbAc8DXgJ8AbgTWXdPwHPiaqv8YuBfuCNmZll/R8COwC7AG8ETomIZ01w3L8A\nPgS8lqqV/VbgLIDM/PNSbO/M3CYzvzhB6G8FDgD2AZ4HvGqCMq8HjgS2rdn/7cDOwMHAB0sc9ToM\neBnwdOCZwHvKuTwXOBX4W+ApwH8C50TEFhGxOfB14PPAk4EvA38zzXH2Bf6X6joeC5xdEslzgD0j\nomvcOX5u/A4ycwC4BDiqXMOjala/qhxjwfjtxouIg4B3A38N7Fj2OTxJ2QXAp6iu005Ur6Fdaoos\nLcd+CVUd3A18ctxuFgHPAhYD/zruXA+i+nK4PXDmNPt7Yzn+blR1sgR4YLrzldQ5TLglNcPXI+K3\nwD3AXwInQdVyDLwOeFdm3puZtwAfpkrsyMz7y+P/AM4Almbm7eP2/d7MfDAzvwOcR5VUj3cYcGpm\n/igzHwTeBbwoIubXGf9rgY9l5u2ZeTdwwgRlTsvM6zJzDdUXgT8Djs7M32fmVcBnqb5M1OsTmbkq\nM38DDAJ9ZfmRwH9m5g8yc21mng48CLyw/G0GfDQzH87MrwCXT3Ocu2rKfxG4ATiwXKcvAocDRMSz\ngfnAuetxDgAfyszfZGY9CeiSUn60XMcPAvtM0sp9MPDNzFyRmQ8B/wrkuH0NlDp7EDgOODjW7R5y\nfGY+kJlXA1cDe9es+35mfj0zHymxT7W/h6kS7WeUOrkyM39Xx/lK6hAm3JKa4VWZuT1V94ajgO9E\nxFjr9GZULcJjbqWmpTIzfwDcBATwpXH7vTsz7xu37c4THH/n2mNk5mrg16zbIjqVnYFVNc9XTVCm\ndtnOwG8y895xsdV7vPH7qz2vPYB/Kl0Xflu+yOxW1u8M3FHzC8DYtlOZqPzYsU4HDo2IoPri86WS\nbK6Pia7VZPYAPlZzXr+hqveJrts6dVK+nP163L6+VrOvUWAtUNvV4/9qHt8PbDNF3FPt7/PAhcBZ\nEXFnRPxbRGxW70lLan8m3JKaprT+nU2VqCwCfkXVOljbgrk7cMfYk4h4G7AFcCfwL+N2+aSI2Hrc\ntndOcOg7a49RtnlK7XGm8XNg15rnE42yUpu03gk8OSK2HRfb2PHuA55Ys+4PJ9hf7TFqz2sVMJiZ\n29f8PTEzh0ucu5QEuXbbqUxU/k6AzLwMeIiqf/ahVInlZLKO5WNfjiY791XA3447t60y89IJ9rtO\nnUTEVlR1WruvA8bta8vMrLfOx5/PpPsrvw4cn5kLgD8FXsH6/Zohqc2ZcEtqmqgcBDwJGM3MtVSt\n1oMRsW3pOvAOqu4jRMQzgQ9QdWt4PfAvEbHPuN0eHxGblz7er6DqtzzeMPCmiNgnIrag6qrwg9KF\nBaqRU542RehfAv4+InaJiO2Bo6c6z8xcBVwKfKjcUPfHVP3PzyhFrgJeHhFPLi39/zDBbt4WEbuW\n/tQDVN07AD4DLImIfcv13DoiDizJ/feBNcDbI2KziPhr4AVTxQo8tab8a4Au4Pya9Z8DPgE8nJlT\njdk93TUkM39J9aXj8KhuRH0zVR/1McuAd5XuK2M31L5mkt19BXhlRPxp6bt+HFVreO2+Bse6o0TE\njuW1t6Em3V9E9EbEc0oXqd9RfYl8ZCOOJanNmHBLaoZvRsRqqmRkkOrGx+vKuqVULZ83ASuALwCn\nlr6xZwAnZubVmfkzqhvqPl+SZqi6BNxN1SJ7JrAkM38y/uCZ+T/Ae4GvUrWMPp2q7/iY44DTS3eB\nifqAfwa4CLgG+DFVQrqGqqV+Mn1UfZ7vBL4GHFvigKql+GrglrLfiW7U/EJZdxPVTY0fKOdyBdVN\nnJ8o534jcERZ9xDVDYdHUHXHOAQ4e4oYAX4A7EX1a8MgcHBm1nbN+DzQzWNfFibzMao+zXdHxMen\nKPdW4J1U3T+eTfXFhBL/14ATqbpm/A5YSXWz6uOU189SqptTfw6spuqPPtbl5WNUN35eFBH3Ut2o\nu+805zCVqfb3h1RfAH5H1dXkO0z9a4CkDhPrdt2TpNkhInqAMzJz1+nKNuDYBwDLMnPSIes2cv+3\nAG+pSdBbpnTVuAt4XvnSs0mKiG2A3wJ7ZebNrY5HkmrZwi1J04iIrSLi5RExNyJ2oRo+72utjqtJ\n/g64fFNMtiPilRHxxNIn/9+Ba6l+NZCkTYqzZ0nS9AI4nqrrxwNUww/+a0sjaoLS0h5MPO74puAg\nqq4bAVwBvC792VbSJsguJZI2eaVbw5eAPwcuyszJbqRrqYgYoerm8tkWx9FDi7rb1MTwbuBpmfmW\nSdYfRtWXf78ZPm7LXyvTndv4GKkmK5rxayFp02ELt6SGi4ik6lt74wbu4mCq8Y6fUiZE0SYuMz84\n9rhMMHQzsNlY/WXmmVQ3us60lr9W6ji3iWJsxLWQtImwD7ek2WAP4Kcm25VxsyVqXRv8WmnidfX1\nLHUYE25JdYmIrogYKUPnXRcRf1WzbiQi3lLz/IiIWFEef7csvjoiVkfEIeuz/4g4nqq/9CFl+/4J\nto2I+EhE3BURv4uIayOiu6w7MCJ+XJaviojjarabHxEZEW8q6+6OiCUR8fyIuKbE8olx5/W9iPhE\nRNwTET+JiMVTXLM3R8Ro2e+FMfEU5bVxHBnVTIU/j4h/rll/XER8JSLOKMPlHRERW0TER0v5O8vj\nLcbt990R8auIuKV0c5gszpGI+FBE/LBcp29ENf732Pq/KnXy21K2q2bd0RFxR0TcGxE3jF2PEvPY\nUIJjr4Hfljp80bjXyKcj4t/HxfSNiHhHebxzRHw1In4ZETdHxNsnOY/HvVYi4gkR8Z6IuLW8Pj4X\nEduNu+79EXEb8O0J9rlDRJxbzv03EXFJRDyhrNstIs4ucf167LVSe251xjhpeUntwYRb0rSimqb6\nm1T9TZ9KNf7xmRHxrOm2zcw/Lw/3zsxtMvNxY05Ptf/MPJZqopovlu2HJjjMflT9YZ8JbAe8lsem\n+b6Pata/7YEDgb+LiPE3Ae5LNRb1IcBHqSaaeSnVONGvjYiXjCv7v1TT0h8LnF2bnNac00FU44b/\nNbAjcAnVBDxT6S1x7AccHREvrVl3ENVYz9tTdT8YAF4I7APsTTXBzXtqyv9hiXEX4I3AKdPU1xuA\nNwM7UY0x/vFyHs8scf9DOY/zqcZV37zs7yjg+Zm5LfAyJh4lZOw1sH2pw++PWz9MlYBGOeaTyjU4\nqyS336Qat3wXYDHwDxHxsvEHmeS1ckT566WamGcbqjHMa72EasKfx+0T+Cfg9nLu86jqNKOa5OZc\n4Faq8dZ3oRoTfEp1vp4ltRkTbkn1eCFVonJCZj6Umd+mSjb6NpH9PwxsC/wR1c3go5n5c4DMHMnM\nazPzkcy8hiq5e8m47d+fmb/PzIuoEvThzLyrTAN+CfDcmrJ3AR8t03l/EbiBKpEfbwnwoRLLGqok\na5/JWrmL4zPzvsy8Fvivcef//cz8ejmPB4DDgPeVOH9JNYrK68ft772Z+WBmfodqZJWJJvUZ8/nM\nXJmZ91FNEvTaklQeApyXmd/KzIepht/bimoK87XAFsCCiNgsM2/JzP+d4hiTuYRqKvUXl+cHl/O9\nE3g+sGNmvq+8Nm6imojodRPv6nEOA/4jM2/KzNXAu4DXxbrdR44r1/2BCbZ/mOpLyB6lzi8pI6G8\nANgZeGfZ9vfTzMQpqYOZcEuqx87Aqsysna76VqpWvfUWEf9dfk5fXbo6rNf+S/eGse1fXBL0TwCf\nBO6KiFMi4g9K2X0jYnn52f8eqkR4h3G7/EXN4wcmeL5NzfM7xg09d2uJf7w9gI+Vrgi/pZr5MSY7\np2LVFPtdNa7szqXMZOXvLsnzdHFOduzNqK7TOscpdbQK2KXcBPsPVDN13hURZ0XEVMeYULmeZ/HY\nF4xDeewmwj2AnceuY7mW76Zqba7HRNdp7rjtx1/bWidRzeZ5UUTcFBHHlOW7AbdO1w+7dOsZe60u\nqzNmSW3GhFtSPe4Edhvru1rsDtxRHt8HPLFm3R9OtbPMPKD8nL5NGdFhuv2P3/7ZNdtfUpZ9PDP/\nBFhA1bXknaX4F6im5N4tM7cDllElvhtql7GuDzVx3jlBuVXA32bm9jV/W2XmpROUHbPbFPsdP4br\nnVTJ6GTlnxTVhDDTxTnZsR+mmu59neOUc9+NUjeZ+YXMXFTKJNXU7OPVM/7sMNXU8HtQddv5alm+\nCrh53HXcNjNfXsc+GR9/Obc1rPulatL4MvPezPynzHwa8FfAO0o/9VXA7jHNjZaZ+cGa1+qSOmOW\n1GZMuCXV4wfA/cC/RMRmUY3z/Eoe67N6FfDXUc369wxg/I2Nv6DqP7uh+59SVDc57lv6gt8H/B4Y\nay3fFvhNZv4+Il5A1Xq6MZ4KvL3E+Rqqvr/nT1BuGfCuiHh2iXG7Un4q7y3X8NnAm6gm2pnMMPCe\niNgxInaguhHvjHFlji99rV8MvAL48hT7OzwiFkTEE4H3AV/JzLVU40UfGBGLy/X9J+BB4NKIeFZE\n/EVUN2v+nurXgEcm2Pcvy/JJXwOZ+WOqBP+zwIWZ+duy6ofAvVHdnLlVRMyJiO6IeP4U51JrGPjH\niNgzqunfx/pP1zVCSES8IiKeUb5o3EPVjeaREtfPgRMiYuuI2DIi/qzOmCR1GBNuSdPKzIeoEuAD\nqJKiTwFvyMyflCIfAR6iSqxP5/FjCh8HnF66BDyuH3Ed+5/OH1D1672bqsvAr6m6AgD8P+B9EXEv\nVVL6pTr3OZkfUN3Y+CtgEDg4M389vlBmfo2qtfesqEYWWUl1flP5DlX3hYuBfy99yifzAarZFa+h\nmtL8R2XZmP+juh53UtXHkmmu5+eB08p2WwJvL+dxA3A4cDLVOb8SeGWpsy2AE8ry/6P6MvKu8TvO\nzPuprtX3ymvghZPE8AWqm1W/ULPtWqovC/tQjeU9lpRvN8W51Dq1nNt3y/a/p7opt157Af8DrAa+\nD3wqM5eXuF4JPAO4jerGyglH4JEkZ5qUpDpFxBHAW0oXipnc73zGTQzTTLGJzJApSe3KFm5JkiSp\ngUy4JUmSpAayS4kkSZLUQLZwS5IkSQ1kwi1JkiQ10JQD9jfSDjvskPPnz2/V4ZvmvvvuY+utt56+\noDZ51mX7sC7bh3XZXqzP9tEJdXnllVf+KjN3rKdsyxLu+fPnc8UVV7Tq8E0zMjJCT09Pq8PQDLAu\n24d12T6sy/ZifbaPTqjLiLi13rJ2KZEkSZIayIRbkiRJaiATbkmSJKmBTLglSZKkBjLhliRJkhrI\nhFuSJElqoGkT7og4NSLuioiVk6yPiPh4RNwYEddExPNmPkxJkiRpdqqnhfs0YP8p1h8A7FX+jgQ+\nvfFhzX7Dw8N0d3ezePFiuru7GR4ebnVI2kDWpSRJ2hjTTnyTmd+NiPlTFDkI+FxmJnBZRGwfETtl\n5s9nKMZZZ3h4mIGBAYaGhli7di1z5syhv78fgL6+vhZHp/VhXUqSpI01E324dwFW1Ty/vSzrWIOD\ngwwNDdHb28vcuXPp7e1laGiIwcHBVoem9WRdSpKkjdXUqd0j4kiqbifMmzePkZGRZh6+aUZHR1m7\ndi0jIyOsXr2akZER1q5dy+joaNuec7uyLtvTWF1q9rMu24v1uWnp7e1tyXGXL1/ekuM20kwk3HcA\nu9U837Use5zMPAU4BWDhwoXZ09MzA4ff9HR1dTFnzhx6enoYGRmhp6eH5cuX09XVRbuec7uyLtvT\nWF1q9rMu24v1uWmpegtvmPnHnMctJxw4g9HMbjORcJ8DHBURZwH7Avd0cv9tgIGBAfr7+x/t97t8\n+XL6+/vthjALDQwMcMghh7D11ltz2223sfvuu3PffffxsY99rNWhSZKkWWLahDsihoEeYIeIuB04\nFtgMIDOXAecDLwduBO4H3tSoYGeLsZvpli5dyujoKF1dXQwODnqT3Sy3Md/0JUlS56pnlJIps8Qy\nOsnbZiyiNtHX10dfX58/j81yg4ODfPGLX6S3t3edLiVLly71C5QkSaqLM01KUxgdHWXRokXrLFu0\naBGjo6MtikiSJM02JtzSFLq6ulixYsU6y1asWEFXV1eLIpIkSbONCbc0hbEbYJcvX86aNWsevQF2\nYGCg1aFJkqRZoqnjcEuzjTfASpKkjWXCLU3DG2AlSdLGsEuJpI4xPDxMd3c3ixcvpru7m+Hh4VaH\nJEnqALZwS+oIw8PDDAwMPDoh1Zw5c+jv7wewi5AkqaFs4ZbUEQYHBxkaGqK3t5e5c+fS29vL0NCQ\nM8BKkhrOhFtSR3BMdUlSq5hwS+oIjqkuSWoVE25JHcEx1SVJreJNk5I6gmOqS5JaxRZuSZIkqYFs\n4ZbUERwWUJLUKrZwS+oIDgsoSWoVE25JHcFhASVJrWLCLakjOCygJKlVTLgldQSHBZQktYo3TUrq\nCA4LKElqFRNuSR2jr6+Pvr4+RkZG6OnpaXU4kqQOYZcSSZIkqYFMuCVJkqQGMuGWpjE8PEx3dzeL\nFy+mu7ub4eHhVockSZJmEftwS1NwdkJJkrSxbOGWpuDshJIkaWOZcEtTcHZCSZK0sUy4pSk4O6Ek\nSdpYJtzSFJydUJIkbSxvmpSm4OyEkiRpY9nCLU2jr6+PlStXcvHFF7Ny5UqT7VnMIR4lSa1gC7ek\njuAQj5KkVrGFW1JHcIhHSVKrmHBL6ggO8ShJahUTbkkdwSEeJUmtYsItqSM4xKMkqVW8aVJSR3CI\nR0lSq5hwS+oYfX199PX1MTIyQk9PT6vDkSR1CLuUSJIkSQ1kwi1JkiQ1kAm3JEmS1EB1JdwRsX9E\n3BARN0bEMROs3y4ivhkRV0fEdRHxppkPVZIkSZp9pk24I2IO8EngAGAB0BcRC8YVextwfWbuDfQA\nH46IzWc4VkmSJGnWqaeF+wXAjZl5U2Y+BJwFHDSuTALbRkQA2wC/AdbMaKSSJEnSLFRPwr0LsKrm\n+e1lWa1PAF3AncC1wN9n5iMzEqEkSZI0i83UONwvA64C/gJ4OvCtiLgkM39XWygijgSOBJg3bx4j\nIyMzdPhN1+rVqzviPDuBddk+rMv2YV22F+uzvViXj6kn4b4D2K3m+a5lWa03ASdkZgI3RsTNwB8B\nP6wtlJmnAKcALFy4MDth4gkn2Ggf1mX7sC7bh3XZXqzPNnLBedZljXq6lFwO7BURe5YbIV8HnDOu\nzG3AYoCImAc8C7hpJgOVJEmSZqNpW7gzc01EHAVcCMwBTs3M6yJiSVm/DHg/cFpEXAsEcHRm/qqB\ncUuSJEmzQl19uDPzfOD8ccuW1Ty+E9hvZkOTJEmSZj9nmpQkSZIayIRbkiRJaiATbkmSJKmBTLgl\nSZKkBjLhliRJkhrIhFuSJElqIBNuSZIkqYFMuCVJkqQGMuGWJEmSGsiEW5IkSWogE25JkiSpgUy4\nJUmSpAYy4ZYkSZIayIRbkiRJaiATbkmSJKmBTLglSZKkBjLhliRJkhrIhFuSJElqIBNuSZIkqYFM\nuCVJkqQGMuGWJEmSGsiEW5IkSWogE25JkiSpgUy4JUmSpAYy4ZYkSZIayIRbkiRJaiATbkmSJKmB\nTLglSZKkBjLhliRJkhrIhFuSJElqIBNuSZIkqYFMuCVJkqQGMuGWJEmSGsiEW5IkSWogE25JkiSp\ngUy4JUmSpAYy4ZYkSZIayIRbkiRJaiATbknSrDM8PEx3dzeLFy+mu7ub4eHhVockSZOa2+oAJEla\nH8PDwwwMDDA0NMTatWuZM2cO/f39APT19bU4Okl6vLpauCNi/4i4ISJujIhjJinTExFXRcR1EfGd\nmQ1TkqTK4OAgQ0ND9Pb2MnfuXHp7exkaGmJwcLDVoUnShKZt4Y6IOcAngb8Ebgcuj4hzMvP6mjLb\nA58C9s/M2yLiqY0KWJLU2UZHR1m0aNE6yxYtWsTo6GiLIpKkqdXTwv0C4MbMvCkzHwLOAg4aV+ZQ\n4OzMvA0gM++a2TAlSap0dXWxYsWKdZatWLGCrq6uFkUkSVOrJ+HeBVhV8/z2sqzWM4EnRcRIRFwZ\nEW+YqQAlSao1MDBAf38/y5cvZ82aNSxfvpz+/n4GBgZaHZokTWimbpqcC/wJsBjYCvh+RFyWmT+t\nLRQRRwJHAsybN4+RkZEZOvyma/Xq1R1xnp3Aumwf1uXsttNOO3HYYYfx5je/mdtuu43dd9+dww8/\nnJ122sl6neV8b7YX6/Ix9STcdwC71TzftSyrdTvw68y8D7gvIr4L7A2sk3Bn5inAKQALFy7Mnp6e\nDQx79hgZGaETzrMTWJftw7qc/Xp6enj/+99vXbYZ67ONXHCedVmjni4llwN7RcSeEbE58DrgnHFl\nvgEsioi5EfFEYF/Au1ckSZLU8aZt4c7MNRFxFHAhMAc4NTOvi4glZf2yzByNiAuAa4BHgM9m5spG\nBi5JkiTNBnX14c7M84Hzxy1bNu75ScBJMxeaJEmSNPs5tbskSZLUQCbckiRJUgOZcEuSJEkNZMIt\nSZIkNZAJtyRJktRAJtySJElSA5lwS5IkSQ1kwi1JkiQ1kAm3JEmS1EAm3JIkSVIDmXBLkiRJDWTC\nLUmSJDWQCbckSZLUQCbckiRJUgOZcEuSJEkNZMItSZIkNZAJt6SOMTw8THd3N4sXL6a7u5vh4eFW\nhyRJ6gBzWx2AJDXD8PAwAwMDDA0NsXbtWubMmUN/fz8AfX19LY5OktTObOGW1BEGBwcZGhqit7eX\nuXPn0tvby9DQEIODg60OTZLU5ky4JXWE0dFRFi1atM6yRYsWMTo62qKIJEmdwi4ldYiIlhw3M1ty\nXKkddXV1sWLFCnp7ex9dtmLFCrq6uloYlSSpE9jCXYfM3OC/PY4+d4O3lTRzBgYG6O/vZ/ny5axZ\ns4bly5fT39/PwMBAq0PTBvAGWEmziS3ckjrC2I2RS5cuZXR0lK6uLgYHB71hchbyBlhJs40t3JI6\nRl9fHytXruTiiy9m5cqVJmezlDfASpptTLglSbOKN8BKmm1MuCVJs8rYDbC1vAFW0qbMhFuSNKt4\nA6yk2cabJiV1jOHhYQYHBx+9aXJgYMB+3LOQN8BKmm1MuCV1BEe2aC99fX309fUxMjJCT09Pq8OR\npCmZcEvqCLUjW4wlaUNDQyxdutSEW1Jb2/v4i7jngYebftz5x5zX1ONtt9VmXH3sfk09Zr1MuCV1\nBEe2kNSp7nngYW454cCmHrMVvz41O8FfH940KakjOLKFJKlVTLgldQRHtpAktYpdSiR1BEe2kCS1\nigm3pI7hyBaSpFawS4kkSZLUQCbckiRJUgOZcEuSJEkNZMItSZIkNZAJtyRJktRAJtySJElSA9WV\ncEfE/hFxQ0TcGBHHTFHu+RGxJiIOnrkQJUmSpNlr2oQ7IuYAnwQOABYAfRGxYJJyJwIXzXSQkiRJ\n0mxVTwv3C4AbM/OmzHwIOAs4aIJyS4GvAnfNYHySJEnSrFbPTJO7AKtqnt8O7FtbICJ2AV4N9ALP\nn2xHEXEkcCTAvHnzGBkZWc9wZ6dOOc92t3r1auuyTViX7cO6bA8XX3wxZ5xxBrfddhu77747hx9+\nOIsXL251WG2l2e+TVr03N9XPg5ma2v2jwNGZ+UhETFooM08BTgFYuHBhdsTUyhec5xTSbcLpwNuH\nddk+rMvZb3h4mDPPPJNTTz2VtWvXMmfOHPr7+1mwYAF9fX2tDq89tCAXacl7cxPOuerpUnIHsFvN\n813LsloLgbMi4hbgYOBTEfGqGYlQkiS1rcHBQYaGhujt7WXu3Ln09vYyNDTE4OBgq0OTZkw9LdyX\nA3tFxJ5UifbrgENrC2TmnmOPI+I04NzM/PoMxilJktrQ6OgoixYtWmfZokWLGB0dbVFE0sybtoU7\nM9cARwEXAqPAlzLzuohYEhFLGh2gJElqX11dXaxYsWKdZStWrKCrq6tFEUkzr64+3Jl5PnD+uGXL\nJil7xMaHJUmSOsHAwAD9/f0MDQ2xdu1ali9fTn9/v11K1FZm6qZJSZKk9TZ2Y+TSpUsZHR2lq6uL\nwcFBb5hUWzHhliRJLdXX10dfX5+jzqht1TW1uyRJkqQNY8ItSZIkNZBdSiTNWlNNtNVImdmS40qS\nZidbuCXNWpm5QX97HH3uBm9rsi1JWl8m3JIkSVIDmXBLkiRJDWTCLUmSJDWQCbckSZLUQCbckiRJ\nUgOZcEvGC/jjAAAUjElEQVSSJEkN5DjckqSWc0x1Se3MFm5JUss5prqkdmbCLUmSJDWQCbckSWqp\n4eFhuru7Wbx4Md3d3QwPD7c6JGlG2YdbkiS1zPDwMAMDAwwNDbF27VrmzJlDf38/AH19fS2OTpoZ\nJtySJKllBgcHOfTQQ1m6dCmjo6N0dXVx6KGHMjg4aMKttmHCLUmSWub666/n/vvvf1wL9y233NLq\n0KQZYx9uSZLUMptvvjlHHXUUvb29zJ07l97eXo466ig233zzVocmzRhbuCVJUss89NBDnHzyyTz3\nuc9l7dq1LF++nJNPPpmHHnqo1aFJM8aEW5IktcyCBQt41atetU4f7sMOO4yvf/3rrQ5NmjEm3JIk\nqWUGBgYmHKVkcHCw1aFJM6ajEu69j7+Iex54uOnHnX/MeU093nZbbcbVx+7X1GNKkrQhxkYiqW3h\ndoQStZuOSrjveeBhbjnhwKYec2RkhJ6enqYes9kJviRJG6Ovr4++vr6W/J8pNYOjlEiSJEkNZMIt\nSZIkNZAJtyRJktRAJtzSNIaHh+nu7mbx4sV0d3czPDzc6pAkqa0sXbqULbfckt7eXrbcckuWLl3a\n6pCkGdVRN01K62t4eHjC4aoA76CXpBmwdOlSli1bxoknnsiCBQu4/vrrOfroowE4+eSTWxydNDNs\n4ZamMDg4yNDQ0DpTDg8NDTk+rCTNkM985jOceOKJvOMd72DLLbfkHe94ByeeeCKf+cxnWh2aNGNM\nuKUpjI6OsmjRonWWLVq0iNHR0RZFJEnt5cEHH2TJkiXrLFuyZAkPPvhgiyKSZp4JtzSFrq4uVqxY\nsc6yFStW0NXV1aKIJKm9bLHFFixbtmydZcuWLWOLLbZoUUTSzLMPtzSFgYEB+vv7H+3DvXz5cqcc\nlqQZ9Na3vvXRPtsLFizgP/7jPzj66KMf1+otzWYm3NIUnHJYkhpr7MbId7/73Tz44INsscUWLFmy\nxBsm1VZMuKVpOOWwJDXWySefzMknn+znbINs23UMzzn9mOYf+PTmHm7bLoADm3vQOplwS5IktbF7\nR0/glhOam4i24svT/GPOa+rx1oc3TUqSJEkNZMItTcOZJiVJ0sawS4k0BWealCRJG6uuFu6I2D8i\nboiIGyPicb3uI+KwiLgmIq6NiEsjYu+ZD1VqPmealCRJG2vahDsi5gCfBA4AFgB9EbFgXLGbgZdk\n5nOA9wOnzHSgUis406QkSdpY9XQpeQFwY2beBBARZwEHAdePFcjMS2vKXwbsOpNBSq0yNtNkb2/v\no8ucaVKSJhcRLTluZrbkuFI96km4dwFW1Ty/Hdh3ivL9wH9PtCIijgSOBJg3bx4jIyP1RTmDmn3M\n1atXd8R5tqtXv/rVHHbYYbzzne9kzz335CMf+QgnnXQS/f39XuNZzvprH9blpmX58uUbvO0RF9zH\naftvvUHb+jqYmvlPa83oTZMR0UuVcC+aaH1mnkLpbrJw4cJs+uD2F5zX9DEhWzKIfwvOs1319PSw\nYMECBgcHH51p8sMf/rA3TM52vkfah3XZXqzPxjD/abl6Eu47gN1qnu9alq0jIv4Y+CxwQGb+embC\nk1rPmSYlSdLGqCfhvhzYKyL2pEq0XwccWlsgInYHzgZen5k/nfEoJUmzwt7HX8Q9Dzzc1GM2e3a5\n7bbajKuP3a+px5Q0u02bcGfmmog4CrgQmAOcmpnXRcSSsn4Z8K/AU4BPlZsl1mTmwsaFLUnaFN3z\nwMNNnULa6aMlzQZ19eHOzPOB88ctW1bz+C3AW2Y2NEmSJGn2c6ZJSS3Vii4IYDcESVLzmHBLaqlm\nd0EAuyFIkpqrrqndJUmSJG0YE25JkiSpgUy4JUmSpAYy4ZYkSZIayIRbkiRJaiATbkmSJKmBOmpY\nwG27juE5px/T/AOf3tzDbdsF0Nxh1iRJkjSxjkq47x09wfF+RUQ0/ZiZ2fRjSpKkTYNdStRxMnOD\n/vY4+twN3laSJHUuE25JkiSpgUy4JUmSpAYy4ZYkSZIayIRbkiRJaiATbkmSJKmBTLglSZKkBjLh\nliRJkhrIhFuSJElqIBNuSZIkqYFMuCVJkqQGMuGWJEmSGmhuqwOQJElSY80/5rzmH/SC5h5zu602\na+rx1ocJtyRJepy9j7+Iex54uOnHbWZiuN1Wm3H1sfs17XitcssJBzb9mPOPOa8lx91UmXBLkqTH\nueeBh5ueMI2MjNDT09O047Wk1VcdyYRbs1IntLxA57S+SJLUzky4NSt1QssL2PoiSVI7cJQSSZIk\nqYFMuCVJkqQGMuGWJEmSGsiEW5IkSWogE25JkiSpgUy4JUmSpAYy4ZYkSZIayIRbkiRJaiATbkmS\nJKmBnGlSUktt23UMzzn9mOYf+PTmHm7bLoDmzo4qSdo0mHBLaql7R0/glhOam4iOjIzQ09PT1GPO\nP+a8ph5PkrTpsEuJJEmS1EAm3JIkSVID1ZVwR8T+EXFDRNwYEY/rbBmVj5f110TE82Y+VEmSJGn2\nmTbhjog5wCeBA4AFQF9ELBhX7ABgr/J3JPDpGY5TkiRJmpXqaeF+AXBjZt6UmQ8BZwEHjStzEPC5\nrFwGbB8RO81wrJIkSdKsU88oJbsAq2qe3w7sW0eZXYCfb1R0DdCSkQIuaO4xt9tqs6YerxUcSk7a\nNLXkven7siE64XO2U+pSrdfUYQEj4kiqLifMmzePkZGRZh6e0/bfeoO26+3tneFI6rN8+fIN3rbZ\n17bZ7h09YYO3vfXEV8xgJPXZ4+hzN2i7rTdr/7qEDf8i3Iq6hA2rz06pyw19b1qXmx7rUhub/8SJ\nG7bdxuQ/m6rIzKkLRLwIOC4zX1aevwsgMz9UU+Y/gZHMHC7PbwB6MnPSFu6FCxfmFVdcsfFnsIlr\nxXi/agzrsn1Yl+3Dumwv1mf76IS6jIgrM3NhPWXr6cN9ObBXROwZEZsDrwPOGVfmHOANZbSSFwL3\nTJVsS5IkSZ1i2i4lmbkmIo4CLgTmAKdm5nURsaSsXwacD7wcuBG4H3hT40KWJEmSZo+6+nBn5vlU\nSXXtsmU1jxN428yGJkmSJM1+zjQpSZIkNZAJtyRJktRAJtySJElSA5lwS5IkSQ1kwi1JkiQ1kAm3\nJEmS1EAm3JIkSVIDTTu1e8MOHPFL4NaWHLy5dgB+1eogNCOsy/ZhXbYP67K9WJ/toxPqco/M3LGe\ngi1LuDtFRFyRmQtbHYc2nnXZPqzL9mFdthfrs31Yl+uyS4kkSZLUQCbckiRJUgOZcDfeKa0OQDPG\numwf1mX7sC7bi/XZPqzLGvbhliRJkhrIFm5JkiSpgUy4Z1hE9ETEn9Y8XxIRb2hlTHq8mayXiHj3\nuOeXzsR+tWko7+lzWx1HO4qIt0fEaEScuQHbvnv6UpJaISIWRsTHJ1n34oi4LiKuiohdIuIrzY6v\nFexSMsMi4jhgdWb+e6tjUXNExOrM3KbVcagxIqIH+OfMfEWrY2k3EfET4KWZefsGbOv7bhMSEUGV\nUzzS6li0aYuIZcCKzDyj1bE0ky3cdYqIr0fEleVb2ZFl2f4R8aOIuDoiLo6I+cAS4B/LN7cXR8Rx\nEfHPpfw+EXFZRFwTEV+LiCeV5SMRcWJE/DAifhoRL27Vec5mk9TR6ogYLHV0WUTMK8tr62UkIj4S\nEVeU1rbnR8TZEfGziPjANPs/Adiq1PeZY8cs/0ZEnBQRKyPi2og4pCzvKcf8SkT8JCLOLP9ZaRoR\n8Yby/rk6Ij4fEfMj4ttl2cURsXspd1pEfLrU+U3lmp9a6ve0mv3tFxHfL+/jL0fENmX5/qVufgT8\ndVn2hPKa2LHm+Y1jz7V+yn+6TwP+OyKOLvXw44i4NCKeVcocUd6LF5Rr/29l+UTvu4nen3PKa2Hs\nPfiPEfH0Uq9jcexV+1z1K++/GyLic8BK4PXlOq+MiBNryq0un4XXRcT/RMQLymfgTRHxVzX7uqS8\nF38U5ZfiqT4vy2f1peXz4IcRsW2p85Mi4vLyufC3rbg27Soito6I88o1XxkRh0xSDxP+MhgRbwFe\nC7y/1OX8iFjZ/DNpgcz0r44/4Mnl362oPljmAauAPcetP46qNYzxz4FrgJeUx+8DPloejwAfLo9f\nDvxPq893Nv5NUEdPARJ4ZVn+b8B7JqiXEeDE8vjvgTuBnYAtgNuBp0y2//J89bg4Vpd//wb4FjCn\nvF5uK/vtAe4BdqX60vt9YFGrr9+m/gc8G/gpsMNYfQDfBN5Ynr8Z+Hp5fBpwFhDAQcDvgOeU630l\nsA/VLGjfBbYu2xwN/CuwZXlv71W2/xJwbilzLPAP5fF+wFdbfV1m8x9wS6mHPwDmlmUvHbuuwBHA\nTcB2pV5uBXYr68a/7yZ6//8J8K2aMtuXf5cD+5THHwSWtvpazMY/YD7wCPBCYOfyGbcjMBf4NvCq\nUi6BA8rjrwEXAZsBewNXleVPBLYsj/cCriiPJ/y8BDYvr43nl3J/UI57JI99zm8BXEH5f9q/Ganz\nvwE+U/N8u0nqoWfsc3OCfZwGHFzzGlrZ6vNqxp8t3PV7e0RcDVwG7Eb1pv5uZt4MkJm/mWrjiNiO\n6sP+O2XR6cCf1xQ5u/x7JdULUOtvfB3tBTwEjH3LnuranlP+vRa4LjN/npkPUn2Q7DbF/qeyCBjO\nzLWZ+QvgO8Dzy7ofZubtWf38etUUcekxfwF8OTN/BY++514EfKGs/zzVNR/zzaw+0a8FfpGZ15br\nfR3V9X4hsAD4XkRcBbwR2AP4I+DmzPxZ2b72Z89TgbG+/28G/mvGz7IzbQd8ubR0fYTqy9WYizPz\nnsz8PXA9VR1NZKL3503A0yLi5IjYn+qLF8BngTdFxBzgEB57DWn93ZqZl1F9to1k5i8zcw1wJo/9\nH/cQcEF5fC3wncx8uDyeX5ZvBnwmIq4Fvkz13hwz0efls4CfZ+blAJn5u3Lc/YA3lPf0D6i+eE33\nWa36XQv8ZVS/yr8Y2J2J60HjzG11ALNBVH04Xwq8KDPvj4gRqjf9H83gYR4s/67Fellvk9TRlsDD\nJWmCqa/t2PV/pObx2PO5U+x/Q9UewzpvjCnrlOq6fysz+2o3ioh9JtthZq6KiF9ExF8ALwAOm9mQ\nO9b7geWZ+eqouuaN1Kyb9r0y2fszM++OiL2Bl1F193st1Relr1L9WvFt4MrM/PUMn08nua+OMrWf\nw4++HzPzkYgYq89/BH5B1er9BOD3Nduvz+dlUP1icWEdcWk9ZeZPI+J5VL/Gf4DqPTSliLiQ6lfe\nKzLzLQ0OcZNlC3d9tgPuLh/kf0TVMrYl8OcRsSdARDy5lL0X2Hb8DjLzHuDueKx/9uupWjw1Myaq\no2bt/+GI2GyCbS4BDil9Cnekau354QzH1Um+DbwmIp4Cj77nLgVeV9YfRnXN63UZ8GcR8Yyyv60j\n4pnAT4D5EfH0Uq5v3HafpWr1/nJmrt2gM9F42wF3lMdH1LlN7ftuwvdnROwAPCEzvwq8B3geQGkt\nvxD4NP5KMVN+CLwkInYovxz0sX7/x21H1VL6CNX/j3OmKX8DsFNEPB+g9BueS1Wvfzf22oiIZ0bE\n1ut5LppEROwM3J/VDY8nAfsycT08KjNflpn7dHKyDbaq1esCYElEjFK9yS8DfknVreTsiHgCcBfw\nl1R9Sr8SEQcBS8ft543Asoh4ItVPnW9qUvydYKI6atb+TwGuiYgfZWZti+fXqLo8XE3Vh/FfMvP/\nSkKg9ZSZ10XEIPCdiFgL/JjqPfZfEfFOqvdk3e+pzPxlRBwBDEfEFmXxe0oLzpHAeRFxP1USX/sl\n+hyqJM1Ebeb8G3B6RLwHOK/ObR5931G1Wk/0/tyF6vUx1rj0rprtzwReTdWfWBspM38eEcdQ9Y8P\n4LzM/MZ67OJTwFejGq71AqZpOc/Mh6K6Ef3kiNgKeIDqV47PUnU5+VG5ufKXwKvW93w0qecAJ0XE\nI8DDwN9R1ff4etA4DgsoSeshIhYCH8lMRxOaxaIapWi7zHxvq2OR1P5s4ZakOpUWvL/DvtuzWkR8\nDXg61Y24ktRwtnBLkiRJDeRNk5IkSVIDmXBLkiRJDWTCLUmSJDWQCbckNUFEbB8R/6/VcUiSms+E\nW5KaY3tgwoR7/EQRm5JNOTZJmi1MuCWpOU4Anh4RV0XESRHRExGXRMQ5wPURMT8iVo4Vjoh/jojj\nyuOnR8QFEXFl2eZxkydFxI4R8a2IuC4iPhsRt5aZFomIwyPih+XY/1lmAiQiVkfEYERcHRGXRcS8\nsvy0iFgWET8A/q3Mwnlq2cePy8ReRMSza/Z7TUTs1eiLKEmzkQm3JDXHMcD/limO31mWPQ/4+8x8\n5jTbngIszcw/Af6Zala+8Y4Fvp2Zzwa+AuwOEBFdwCHAn2XmPsBaHhtHfGvgsszcG/gu8Naa/e0K\n/GlmvgMYKPt+AdBLNdPc1sAS4GNlvwuB2+u8FpLUUfypUJJa54eZefNUBSJiG+BPgS9XM1UDsMUE\nRRdRTVVOZl4QEXeX5YuBPwEuL9tvBdxV1j0EnFseXwn8Zc3+vpyZa8vj/YC/KrMzAmxJldB/HxiI\niF2BszPzZ1OfriR1JhNuSWqd+2oer2HdXx23LP8+AfhtaUXeEAGcnpnvmmDdw/nY7GdrWff/hNrY\nAvibzLxh3PajpdvJgcD5EfG3mfntDYxTktqWXUokqTnuBbadYv0vgKdGxFMiYgvgFQCZ+Tvg5oh4\nDUBU9p5g++8Bry1l9gOeVJZfDBwcEU8t654cEXusZ+wXAkujNJFHxHPLv08DbsrMjwPfAP54Pfcr\nSR3BhFuSmiAzfw18LyJWRsRJE6x/GHgf8EPgW8BPalYfBvRHxNXAdcBBExzieGC/cuPla4D/A+7N\nzOuB9wAXRcQ1Zd87rWf47wc2A66JiOvKc6gS/JURcRXQDXxuPfcrSR0hHvs1UZI0W5VW8bWZuSYi\nXgR8eiO6oUiSZpB9uCWpPewOfCkinkB1M+RbpykvSWoSW7glSZKkBrIPtyRJktRAJtySJElSA5lw\nS5IkSQ1kwi1JkiQ1kAm3JEmS1EAm3JIkSVID/X/O4kbs/oUSeAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d8e64c2b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tag_index = 0\n",
    "col_name = \"out-of-sample prob positive for \" + my_tags[tag_index]\n",
    "probpos = pd.DataFrame({col_name:probs[[tag_index]].sum(axis=1), \n",
    "                        \"true genres\": [r['y'] for r in revtest]})\n",
    "probpos.boxplot(col_name,by=\"true genres\", figsize=(12,5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The chart above summarizes the probability of each plot being a scifi movie based on the reviews"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>out-of-sample prob positive for sci-fi</th>\n",
       "      <th>true genres</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>doc</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.140609</td>\n",
       "      <td>comedy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.025672</td>\n",
       "      <td>romance</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     out-of-sample prob positive for sci-fi true genres\n",
       "doc                                                    \n",
       "0                                  0.140609      comedy\n",
       "1                                  0.025672     romance"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "probpos[0:2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "target = [r['y'] for r in revtest]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.341563786008\n",
      "confusion matrix\n",
      " [[14  2  6  5  7  8]\n",
      " [ 4  7  3  7  6  4]\n",
      " [11  9 36  7 18  5]\n",
      " [ 8  2  1  2  1  2]\n",
      " [ 6  4  9  5 10  1]\n",
      " [ 6  4  5  1  3 14]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzoAAALWCAYAAACOQOb5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRCciIiIiIrpOEp2IiIiIiOg6/w/Ly2DTWt0dkgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d8c2dbb70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.evaluate_prediction(predictions, target, my_tags, \"Deep IR with word2vec\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# To Be continued"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise:** Can you explain why we got very bad results for deepir? (Hint: training dataset contains a total of 30K words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise:** Feel free to test other approaches proposed by the original authors [here](https://github.com/RaRe-Technologies/movie-plots-by-genre/blob/master/Document%20classification%20with%20word%20embeddings%20tutorial.ipynb)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " **Excercise:** Experiment with other type of classifiers, like Random Forest Classifiers and Gradient Boost, and if possible, Deep Neural Networks. [More information](http://scikit-learn.org/stable/supervised_learning.html)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise:** Try to implement different preprocessing methods such as stopwords removal, and those other ones you learned in the Text Mining course"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Excercise:** Create better visualization (always good to know your data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}