Emotion_Recognition_Word_Embeddings.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Word Embeddings for Text Classification (Emotion Recognition)\n",
    "\n",
    "**Task:** Classify text data into different emotions using word embedding and other deep information retrieval approaches.\n",
    "\n",
    "**Github Repo:** [LINK](https://github.com/omarsar/data_mining_lab_fall_2)\n",
    "\n",
    "**Dataset: ** [SemEval 2017 Task](https://competitions.codalab.org/competitions/16380)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "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": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-11-19 21:47:52,626 : INFO : 'pattern' package not found; tag filters are not available for English\n",
      "/home/ellfae/anaconda3/envs/tensorflow/lib/python3.6/site-packages/sklearn/cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "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"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "### train data\n",
    "anger_train = pd.read_csv(\"data/semeval/train/anger-ratings-0to1.train.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "sadness_train = pd.read_csv(\"data/semeval/train/sadness-ratings-0to1.train.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "fear_train = pd.read_csv(\"data/semeval/train/fear-ratings-0to1.train.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "joy_train = pd.read_csv(\"data/semeval/train/joy-ratings-0to1.train.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "train_data = pd.concat([anger_train, fear_train, joy_train, sadness_train])\n",
    "\n",
    "### test data\n",
    "anger_test = pd.read_csv(\"data/semeval/dev/anger-ratings-0to1.dev.gold.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "sadness_test = pd.read_csv(\"data/semeval/dev/sadness-ratings-0to1.dev.gold.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "fear_test = pd.read_csv(\"data/semeval/dev/fear-ratings-0to1.dev.gold.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "joy_test = pd.read_csv(\"data/semeval/dev/joy-ratings-0to1.dev.gold.txt\",\n",
    "                         sep=\"\\t\", header=None, names=[\"id\", \"text\", \"emotion\", \"intensity\"])\n",
    "test_data = pd.concat([anger_test, fear_test, joy_test, sadness_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# shuffle dataset\n",
    "train_data = train_data.sample(frac=1)\n",
    "test_data = test_data.sample(frac=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>id</th>\n",
       "      <th>text</th>\n",
       "      <th>emotion</th>\n",
       "      <th>intensity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>617</th>\n",
       "      <td>20617</td>\n",
       "      <td>Recording some more #FNAF and had to FaceTime ...</td>\n",
       "      <td>fear</td>\n",
       "      <td>0.458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>992</th>\n",
       "      <td>20992</td>\n",
       "      <td>@darwinwatersons @pennyfitzger31 @gumballwatte...</td>\n",
       "      <td>fear</td>\n",
       "      <td>0.271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>144</th>\n",
       "      <td>20144</td>\n",
       "      <td>@Budget  car rental you have made realize why ...</td>\n",
       "      <td>fear</td>\n",
       "      <td>0.729</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>224</th>\n",
       "      <td>20224</td>\n",
       "      <td>Retweeted Dr. Rand Paul (@RandPaul):\\n\\nStop f...</td>\n",
       "      <td>fear</td>\n",
       "      <td>0.667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>385</th>\n",
       "      <td>40385</td>\n",
       "      <td>@SimonSSSJ123 @EllieG10853 @Onision @Eugenia_C...</td>\n",
       "      <td>sadness</td>\n",
       "      <td>0.485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>574</th>\n",
       "      <td>10574</td>\n",
       "      <td>@MMASOCCERFAN @outmagazine No offense but the ...</td>\n",
       "      <td>anger</td>\n",
       "      <td>0.417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>281</th>\n",
       "      <td>10281</td>\n",
       "      <td>Have wee pop socks on and they KEEP FALLING OF...</td>\n",
       "      <td>anger</td>\n",
       "      <td>0.562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>579</th>\n",
       "      <td>30579</td>\n",
       "      <td>@Devilligan It's a beautifully sincere balanci...</td>\n",
       "      <td>joy</td>\n",
       "      <td>0.375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>609</th>\n",
       "      <td>10609</td>\n",
       "      <td>I've been wanting salty fries from McDonald's ...</td>\n",
       "      <td>anger</td>\n",
       "      <td>0.396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>231</th>\n",
       "      <td>30231</td>\n",
       "      <td>Ryan Gosling and Eva Mendes finally  ; B joyfu...</td>\n",
       "      <td>joy</td>\n",
       "      <td>0.620</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        id                                               text  emotion  \\\n",
       "617  20617  Recording some more #FNAF and had to FaceTime ...     fear   \n",
       "992  20992  @darwinwatersons @pennyfitzger31 @gumballwatte...     fear   \n",
       "144  20144  @Budget  car rental you have made realize why ...     fear   \n",
       "224  20224  Retweeted Dr. Rand Paul (@RandPaul):\\n\\nStop f...     fear   \n",
       "385  40385  @SimonSSSJ123 @EllieG10853 @Onision @Eugenia_C...  sadness   \n",
       "574  10574  @MMASOCCERFAN @outmagazine No offense but the ...    anger   \n",
       "281  10281  Have wee pop socks on and they KEEP FALLING OF...    anger   \n",
       "579  30579  @Devilligan It's a beautifully sincere balanci...      joy   \n",
       "609  10609  I've been wanting salty fries from McDonald's ...    anger   \n",
       "231  30231  Ryan Gosling and Eva Mendes finally  ; B joyfu...      joy   \n",
       "\n",
       "     intensity  \n",
       "617      0.458  \n",
       "992      0.271  \n",
       "144      0.729  \n",
       "224      0.667  \n",
       "385      0.485  \n",
       "574      0.417  \n",
       "281      0.562  \n",
       "579      0.375  \n",
       "609      0.396  \n",
       "231      0.620  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "58792"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Count the number of words in training dataset\n",
    "train_data['text'].apply(lambda x: len(x.split(' '))).sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note:** Usually at least 500K words are suggested to train a word2vec model, so performance may not be the best here."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1:** Plot word frequency in both train and test dataset. (Hint: refer to DM lab 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fed1a136cf8>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAD8CAYAAACRkhiPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nApLUMUNAkjpmCEhSxwwBSerY/wdBLzj6zJk7XQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fed1a0e45f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### classes / targets\n",
    "my_tags = ['sadness', 'fear', 'joy', 'anger']\n",
    "train_data.emotion.value_counts().plot(kind=\"bar\", rot=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Is this considered an imbalanced dataset? Try to perform some more data exploration and statistical inference as done in lab session 1. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fed19feb320>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/Y80AYwvAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fed19ff0ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_data.emotion.value_counts().plot(kind=\"bar\", rot=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "### Train Naive Models (Baselines)\n",
    "- Bag of words\n",
    "- N-grams\n",
    "- TF-IDF"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bag of Words\n",
    "Using scikit-learn CountVectorizer perform word frequency and use these as features to train a model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# build analyzers\n",
    "count_vectorizer = CountVectorizer(\n",
    "    analyzer=\"word\", tokenizer=nltk.word_tokenize,\n",
    "    preprocessor=None, stop_words='english', max_features=3000) \n",
    "\n",
    "# apply analyzer\n",
    "train_data_features = count_vectorizer.fit_transform(train_data['text'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 2, 0, ..., 0, 0, 0],\n",
       "       [1, 0, 0, ..., 0, 0, 0],\n",
       "       [6, 2, 0, ..., 0, 0, 0],\n",
       "       ..., \n",
       "       [1, 3, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [1, 2, 0, ..., 0, 0, 0]], dtype=int64)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data_features.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "logreg_model = linear_model.LogisticRegression(n_jobs=1, C=1e5)\n",
    "logreg_model = logreg_model.fit(train_data_features, train_data['emotion'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['af',\n",
       " 'affair',\n",
       " 'affliction',\n",
       " 'afford',\n",
       " 'afraid',\n",
       " 'afternoon',\n",
       " 'age',\n",
       " 'agent',\n",
       " 'ages',\n",
       " 'aggravate']"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# observe some features\n",
    "count_vectorizer.get_feature_names()[100:110]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.734870317003\n",
      "confusion matrix\n",
      " [[60  6  3 15]\n",
      " [10 80  6 14]\n",
      " [ 5  9 65  0]\n",
      " [ 5 15  4 50]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAEmCAYAAADbUaM7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fed19a97400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.predict(count_vectorizer, logreg_model, test_data, my_tags)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# words for the emotion of anger\n",
    "anger_tag_id = word_embeddings.get_tag_index(my_tags, \"anger\") \n",
    "anger_words = word_embeddings.most_influential_words(logreg_model, count_vectorizer, \\\n",
    "                                                      anger_tag_id, 3000)\n",
    "anger_words = pd.DataFrame(anger_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>sober</td>\n",
       "      <td>32.652855</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>grim</td>\n",
       "      <td>29.568752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>spent</td>\n",
       "      <td>29.535102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>stayed</td>\n",
       "      <td>28.740246</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>sunk</td>\n",
       "      <td>28.026766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>pine</td>\n",
       "      <td>27.139517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>depress</td>\n",
       "      <td>26.822598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>depression</td>\n",
       "      <td>26.498302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>unhappy</td>\n",
       "      <td>26.318313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>sadly</td>\n",
       "      <td>26.235318</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            0          1\n",
       "0       sober  32.652855\n",
       "1        grim  29.568752\n",
       "2       spent  29.535102\n",
       "3      stayed  28.740246\n",
       "4        sunk  28.026766\n",
       "5        pine  27.139517\n",
       "6     depress  26.822598\n",
       "7  depression  26.498302\n",
       "8     unhappy  26.318313\n",
       "9       sadly  26.235318"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "anger_words[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# words for the emotion of fear\n",
    "fear_tag_id = word_embeddings.get_tag_index(my_tags, \"fear\") \n",
    "fear_words = word_embeddings.most_influential_words(logreg_model, count_vectorizer, \\\n",
    "                                                      fear_tag_id, 3000)\n",
    "fear_words = pd.DataFrame(fear_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>horror</td>\n",
       "      <td>29.322370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>concern</td>\n",
       "      <td>28.709943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>flirting</td>\n",
       "      <td>27.319602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>shake</td>\n",
       "      <td>25.605281</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>terrorism</td>\n",
       "      <td>25.560661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>nervous</td>\n",
       "      <td>25.427409</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>horrific</td>\n",
       "      <td>23.865957</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>bully</td>\n",
       "      <td>23.771854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>jimmy</td>\n",
       "      <td>23.579379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>rooney</td>\n",
       "      <td>23.255897</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0          1\n",
       "0     horror  29.322370\n",
       "1    concern  28.709943\n",
       "2   flirting  27.319602\n",
       "3      shake  25.605281\n",
       "4  terrorism  25.560661\n",
       "5    nervous  25.427409\n",
       "6   horrific  23.865957\n",
       "7      bully  23.771854\n",
       "8      jimmy  23.579379\n",
       "9     rooney  23.255897"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fear_words[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[['rich', 27],\n",
       "  ['idea', 208],\n",
       "  ['😭', 283],\n",
       "  ['cold', 332],\n",
       "  ['finish', 353],\n",
       "  ['honestly', 360],\n",
       "  ['plenty', 375],\n",
       "  ['bake', 383],\n",
       "  ['childhood', 405],\n",
       "  ['sue', 447],\n",
       "  ['definitely', 475],\n",
       "  ['father', 480],\n",
       "  ['birth', 483],\n",
       "  ['red', 488]],\n",
       " 14]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# check overlap of words and their index\n",
    "word_embeddings.check_word_overlap(anger_words[:500][0], fear_words[:500][0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You may notice that even though the words overlap some of them are not even related to any of the emotions, therefore, we can improve this model. What would happen if we just remove the overlapping words?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Character N-grams\n",
    "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": 28,
   "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['text'])\n",
    "\n",
    "charn_model = charn_model.fit(train_data_features, train_data['emotion'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[' bec',\n",
       " ' beca',\n",
       " ' bee',\n",
       " ' been',\n",
       " ' bei',\n",
       " ' bein',\n",
       " ' bet',\n",
       " ' bi',\n",
       " ' bl',\n",
       " ' bla']"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### observe some features\n",
    "n_gram_vectorizer.get_feature_names()[100:110]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.662824207493\n",
      "confusion matrix\n",
      " [[58  4  9 13]\n",
      " [14 71  6 19]\n",
      " [10 15 49  5]\n",
      " [ 7  9  6 52]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAEmCAYAAADbUaM7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fed0b1760f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.predict(n_gram_vectorizer, charn_model, test_data, my_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can notice that the char n-gram model doesn't perform as well as the bag of words model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Try to improve the model by removing stopwords; or try other techniques you have learned so far."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### TF-IDF\n",
    "Term Frequency - Inverse Document Frequency is useful to rank importance of words to documents"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "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['text'])\n",
    "\n",
    "tfidf_model = linear_model.LogisticRegression(n_jobs=1, C=1e5)\n",
    "tfidf_model = tfidf_model.fit(train_data_features, train_data['emotion'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['17',\n",
       " '18:41',\n",
       " '1nataliemaines',\n",
       " '1st',\n",
       " '2',\n",
       " '2-3',\n",
       " '20',\n",
       " '2014',\n",
       " '2015.',\n",
       " '2016']"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf_vect.get_feature_names()[100:110]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.798270893372\n",
      "confusion matrix\n",
      " [[62  7  3 12]\n",
      " [ 6 89  5 10]\n",
      " [ 4  4 69  2]\n",
      " [ 6  8  3 57]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAEmCAYAAADbUaM7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fed0b6c4ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_embeddings.predict(tf_vect, tfidf_model, test_data, my_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us perform some further analysis on the TF-IDF model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['unhappy', 49.125246254034337],\n",
       " ['sober', 48.022596096955574],\n",
       " ['lost', 46.845720591869181],\n",
       " ['pine', 45.987331531830989],\n",
       " ['blues', 45.952139717565011],\n",
       " ['grim', 45.896981357209285],\n",
       " ['stayed', 45.417541326629191],\n",
       " ['dark', 44.438956045973413],\n",
       " ['dull', 43.812737961256708],\n",
       " ['depressing', 43.002286472820778]]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# words for the anger emotion\n",
    "anger_tag_id = word_embeddings.get_tag_index(my_tags, \"anger\")\n",
    "anger_words = word_embeddings.most_influential_words(tfidf_model, tf_vect, anger_tag_id, 3000)\n",
    "anger_words[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['elated', 49.435472553063647],\n",
       " ['pleasing', 42.330530868032966],\n",
       " ['playful', 42.251062577888234],\n",
       " ['animated', 41.807040398506309],\n",
       " ['joyous', 40.796975736039876],\n",
       " ['rejoicing', 40.583376648348029],\n",
       " ['optimism', 39.74570894231055],\n",
       " ['hilarious', 39.619336751035114],\n",
       " ['cheerfully', 38.552804700392308],\n",
       " ['cheering', 38.203363194562726]]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# words for the joy\n",
    "joy_tag_id = word_embeddings.get_tag_index(my_tags, \"joy\")\n",
    "joy_words = word_embeddings.most_influential_words(tfidf_model, tf_vect, joy_tag_id, 3000)\n",
    "joy_words[0:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Head to scikit-learn webiste and learn how to use different parameters for the TF-IDF vectorizer, such as modifying the ngram_range. It might improve the performance of the classifier."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Averaging Word Vectors (word2vec model)\n",
    "\n",
    "**SOURCE:** https://code.google.com/archive/p/word2vec/\n",
    "\n",
    "First, we are going to use a pretrained model provided as open source by Google (word2vec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-11-19 22:38:33,997 : INFO : loading projection weights from data/google_word2vec/GoogleNews-vectors-negative300.bin.gz\n",
      "2017-11-19 22:40:19,543 : INFO : loaded (3000000, 300) matrix from data/google_word2vec/GoogleNews-vectors-negative300.bin.gz\n",
      "2017-11-19 22:40:19,543 : 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": [
    "**Pre-processing** 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": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_tokenized = test_data.apply(lambda r: word2vec_helpers.w2v_tokenize_text(r['text']), axis=1).values\n",
    "train_tokenized = train_data.apply(lambda r: word2vec_helpers.w2v_tokenize_text(r['text']), axis=1).values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ list(['Dunno', 'am', 'going', 'to', 'the', 'Yorkshire', 'scare', 'grounds', 'when', 'only', 'lasted', 'minute', 'in', 'the', 'Alton', 'towers', 'one', 'before', 'running', 'out', 'fire', 'exit', 'crying']),\n",
       "       list(['Ananya', 'just', 'grabbed', 'bible', 'opened', 'it', 'started', 'reading', 'and', 'then', 'said', \"'where\", 'do', 'they', 'talk', 'about', 'burning', 'people']),\n",
       "       list(['kayleighmcenany', 'DonaldJTrumpJr', 'Is', 'that', 'really', 'all', 'you', 'can', 'offer', 'for', 'those', 'who', 'sacrifice', 'daily', 'to', 'keep', 'you', 'safe', '...', 'kayleighmcenany']),\n",
       "       list(['incredible', 'that', 'anthony', 'weiner', 'has', 'been', 'caught', 'chasing', 'busty', 'milkmaid', 'across', 'park', 'at', 'double', 'speed', 'while', 'jaunty', 'music', 'plays']),\n",
       "       list(['This', 'is', 'horrible', 'Lewis', 'Dunk', 'has', 'begun', 'networking', 'Neo-Geo', 'with', 'his', 'holiday', 'home', 'in', 'Mexico']),\n",
       "       list(['Just', 'seeing', 'Alex', 'revells', 'face', 'gets', 'me', 'angry']),\n",
       "       list(['harrietemmett', 'great', 'minds', 'think', 'alike']),\n",
       "       list(['CesarSampao', 'thisisbolton', 'do', \"n't\", 'get', 'me', 'started', 'on', 'town', 'centre', 'Used', 'to', 'go', 'every', 'week', '...', 'not', 'been', 'for', '18', 'months']),\n",
       "       list(['Really', '...', '..', 'Jumanji', '...', '.w/', 'The', 'Rock', 'Jack', 'Black', 'and', 'Kevin', 'Hart', '...', 'are', 'you', 'kidding', 'me', 'WTF', 'ThisIsATerribleIdea']),\n",
       "       list(['Gronnhair', 'buryprofs', 'DittoBistro', 'it', 'was', 'indeed', 'lovely', 'and', 'the', 'team', 'were', 'incredibly', 'attentive', 'and', 'on', 'the', 'ball', 'Cheese', 'was', 'lively', 'gesture'])], dtype=object)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_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 converted into its vector representation obtained from word2vec). Since we are using a pretrained word2vec the conversion can be done easily. \n",
    "\n",
    "Now that we have the word embeddings, we can obtain vectors for an entire document, by averaging word vectors. (Warning: very naive approach!)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![alt txt](https://docs.google.com/drawings/d/1m1Uod7QU4iSN7J3iqsTU6JRWdGTEJRZ2dA7rZrgNHgA/pub?w=942&h=375)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Here we average the word vectors\n",
    "X_train_word_average = word2vec_helpers.word_averaging_list(wv,train_tokenized)\n",
    "X_test_word_average = word2vec_helpers.word_averaging_list(wv,test_tokenized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.0828345 , -0.03720188,  0.03113783,  0.09435879, -0.04738354,\n",
       "       -0.02334516,  0.04464549, -0.08764581,  0.05910314,  0.05702304,\n",
       "       -0.0381741 , -0.17588666, -0.00759566, -0.01538563, -0.06418473,\n",
       "        0.0447026 ,  0.10272776,  0.07494403, -0.06186917, -0.07438852,\n",
       "       -0.00160431,  0.09135552,  0.05929718, -0.00096006,  0.03247862,\n",
       "       -0.01421031, -0.03617261,  0.03098794,  0.06477017, -0.03398748,\n",
       "        0.02085624,  0.05060841, -0.05719795, -0.02933487, -0.04135398,\n",
       "       -0.04650778,  0.03750284, -0.04152595,  0.01536915,  0.08942927,\n",
       "        0.09623399, -0.11036112,  0.1455759 , -0.02299988, -0.00287363,\n",
       "        0.05799081, -0.05199031, -0.01896928,  0.08540779,  0.01586382,\n",
       "       -0.02355479,  0.03279874,  0.00465961,  0.03000174, -0.03700826,\n",
       "        0.0488888 , -0.04989709, -0.06193588,  0.06018326, -0.08827129,\n",
       "       -0.02334823,  0.04257724, -0.05347706, -0.04378926, -0.05149174,\n",
       "        0.00481888, -0.04781246,  0.08248097, -0.01080719,  0.06393816,\n",
       "        0.01756501,  0.07793213,  0.07762833,  0.08007104, -0.15161249,\n",
       "       -0.03722691,  0.12395981,  0.06855031,  0.0636224 ,  0.09780615,\n",
       "       -0.0548266 , -0.05845098,  0.04592308, -0.02027553, -0.07172428,\n",
       "       -0.01303808, -0.04026437,  0.13182129, -0.03476157, -0.00912584,\n",
       "        0.03185095,  0.06501495, -0.06281661, -0.03742494, -0.03443837,\n",
       "       -0.0775518 ,  0.05683483,  0.07658903, -0.00182815, -0.02445133,\n",
       "       -0.05789818, -0.10756478, -0.0201588 , -0.01367355, -0.02246436,\n",
       "       -0.00159421, -0.00341015, -0.01681556,  0.06667787, -0.04837777,\n",
       "       -0.07107475,  0.01787926, -0.01537318, -0.03864638,  0.02526326,\n",
       "        0.00769131,  0.0201249 , -0.03280054,  0.04744314,  0.04775037,\n",
       "       -0.11188859,  0.0657398 , -0.07222334,  0.04976749, -0.01304947,\n",
       "       -0.073121  , -0.06879081, -0.0513772 ,  0.00881809, -0.03296001,\n",
       "       -0.06841825, -0.06736603, -0.07046579, -0.01661756, -0.02274008,\n",
       "       -0.03076619,  0.08739944,  0.03017626, -0.00196825,  0.01702739,\n",
       "        0.03925883, -0.05889527,  0.01752211, -0.06921595,  0.00865561,\n",
       "        0.08357184, -0.09631499, -0.0878009 , -0.0602692 , -0.03293969,\n",
       "        0.0819206 ,  0.04106344, -0.06630815,  0.06256788, -0.03175091,\n",
       "        0.06215457, -0.05327656, -0.06572698, -0.04440877,  0.00412845,\n",
       "       -0.0150367 ,  0.03353119,  0.06293762,  0.07193726,  0.00107012,\n",
       "       -0.13672085,  0.04287213, -0.03575428,  0.06815045, -0.05873841,\n",
       "       -0.10887971,  0.00981752, -0.03836062, -0.08378714, -0.06903499,\n",
       "       -0.05881749,  0.09471442, -0.04757425, -0.01046868,  0.04602438,\n",
       "       -0.07361519,  0.01670819,  0.10008384,  0.04938151,  0.02994714,\n",
       "        0.00915309, -0.09731684,  0.00914007,  0.04141181,  0.01284424,\n",
       "        0.01567608,  0.01832188,  0.02425939,  0.00986745, -0.06187043,\n",
       "        0.04872014, -0.00831913, -0.03874445, -0.05126166, -0.12295007,\n",
       "       -0.00791001,  0.02206907, -0.042601  ,  0.02246608,  0.05571443,\n",
       "       -0.03746784, -0.12760188,  0.00159992,  0.06182094, -0.04096721,\n",
       "       -0.00665803,  0.00154519,  0.01717271,  0.01050951, -0.07053998,\n",
       "       -0.00748874,  0.08824413, -0.02599456, -0.12767237, -0.05009941,\n",
       "       -0.00453471,  0.08546905,  0.01298815, -0.02659122,  0.09782809,\n",
       "       -0.07475995,  0.0382194 ,  0.06934613, -0.04686676, -0.00734404,\n",
       "        0.01479908, -0.07444905,  0.01112435, -0.02259457,  0.04467384,\n",
       "        0.02088736, -0.01236321, -0.09502137,  0.03587033,  0.02884967,\n",
       "        0.03337711,  0.02563795,  0.00632766, -0.08291794, -0.02519703,\n",
       "        0.04240022,  0.0151681 ,  0.09008405,  0.00053487,  0.00060616,\n",
       "        0.03134907, -0.00020461,  0.08677844,  0.04045636,  0.04757625,\n",
       "        0.01134111,  0.01996336,  0.02787027, -0.04714117, -0.05198164,\n",
       "       -0.02500892,  0.02221724, -0.079062  ,  0.01837583,  0.03457611,\n",
       "        0.13048233, -0.06731659,  0.020221  , -0.05894107, -0.00918363,\n",
       "        0.06769336,  0.09033087,  0.13428234,  0.02842295,  0.09028544,\n",
       "       -0.06185293, -0.04030589, -0.11185527, -0.01389695, -0.05867941,\n",
       "        0.04924538, -0.01998544, -0.00255985,  0.08157858,  0.03989211,\n",
       "        0.02309393, -0.05822608, -0.02368595,  0.02993208,  0.0644683 ,\n",
       "       -0.05595275,  0.05409642, -0.11405333,  0.00312885, -0.10418902,\n",
       "       -0.0367443 ,  0.00490854, -0.03130596,  0.03907832, -0.00813656], dtype=float32)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_word_average[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Further exploration of word2vec generated vectors**\n",
    "\n",
    "Some tips here: https://radimrehurek.com/gensim/models/keyedvectors.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  7.44407773e-02,   5.99248260e-02,  -1.14452699e-02,\n",
       "        -9.60286036e-02,  -6.58800900e-02,   8.22570622e-02,\n",
       "         3.55454721e-02,  -1.34923914e-02,   6.36468679e-02,\n",
       "         9.82618257e-02,  -4.52227741e-02,  -9.44467355e-03,\n",
       "        -3.59176770e-02,   4.80143018e-02,  -3.16373296e-02,\n",
       "         5.06197289e-02,   7.18353540e-02,   1.92615520e-02,\n",
       "         2.38210484e-02,  -8.88636801e-03,  -2.18669791e-02,\n",
       "         8.93289372e-02,   1.46090025e-02,   1.07008619e-02,\n",
       "         1.19105242e-01,   4.76420969e-02,   7.58365449e-03,\n",
       "         1.37715442e-02,   9.04455483e-02,  -5.84360100e-02,\n",
       "         6.21580519e-02,  -3.23817395e-02,   5.54583818e-02,\n",
       "         9.49119963e-03,   1.11661166e-01,  -8.33736733e-02,\n",
       "         9.67730135e-02,   3.25678401e-02,  -3.85231040e-02,\n",
       "         4.24312428e-02,   3.77786942e-02,  -1.43298497e-02,\n",
       "        -7.67670525e-03,  -8.52346942e-02,  -5.21085449e-02,\n",
       "        -1.04961500e-01,  -4.42922637e-02,   3.59176770e-02,\n",
       "        -7.22075552e-02,  -5.95526211e-03,  -7.36963674e-02,\n",
       "        -1.39576457e-02,  -2.73569860e-02,  -1.12126423e-02,\n",
       "        -1.58186648e-02,   2.47515589e-02,  -1.17709478e-02,\n",
       "        -4.24312428e-02,   5.69471940e-02,  -8.85845274e-02,\n",
       "         4.59671803e-02,   9.97506455e-02,  -1.01983868e-01,\n",
       "         2.40071509e-02,  -6.69967011e-02,  -1.97268073e-02,\n",
       "        -1.98198576e-02,   7.02534849e-03,  -1.47951050e-02,\n",
       "        -6.68803870e-04,  -1.95407048e-02,  -4.35478538e-02,\n",
       "         1.74935833e-02,   6.77411109e-02,  -1.39204249e-01,\n",
       "         3.46614886e-03,   1.35854417e-02,   9.52841938e-02,\n",
       "         2.38210484e-02,   1.24316104e-01,   4.03841212e-02,\n",
       "         1.24316104e-01,   1.98198576e-02,   1.07939132e-01,\n",
       "         8.70026555e-03,  -7.33241662e-02,  -6.62522912e-02,\n",
       "         1.56325642e-02,  -1.26549322e-02,   4.20590416e-02,\n",
       "         1.28410347e-02,   4.03841212e-02,  -6.62522912e-02,\n",
       "        -1.85171440e-02,  -5.44348173e-03,   1.74935833e-02,\n",
       "         3.36844511e-02,  -6.21580519e-02,  -5.91804199e-02,\n",
       "         6.79272087e-03,   6.84855133e-02,   1.22827284e-01,\n",
       "        -6.25302568e-02,   7.10909441e-02,   5.13641387e-02,\n",
       "        -1.58186648e-02,   1.37715444e-01,  -6.40190691e-02,\n",
       "         3.74064930e-02,  -4.87587117e-02,   6.81133121e-02,\n",
       "        -7.63017982e-02,   4.01980206e-02,  -3.12651284e-02,\n",
       "         3.74064930e-02,  -5.69471940e-02,   1.87590763e-01,\n",
       "        -1.45903930e-01,   4.57810797e-02,   3.21956351e-02,\n",
       "        -9.15621594e-02,   4.48505692e-02,   8.04890878e-03,\n",
       "        -1.42367985e-02,  -4.70837913e-02,  -1.21338472e-01,\n",
       "        -2.75430884e-02,  -8.93289372e-02,  -4.31756526e-02,\n",
       "        -4.41061631e-02,  -4.15007323e-02,   1.50742577e-02,\n",
       "        -8.03960413e-02,   1.12405576e-01,  -3.65225063e-03,\n",
       "        -2.67986804e-02,  -5.91804199e-02,   1.87590763e-01,\n",
       "        -6.81133121e-02,   4.76420969e-02,  -3.10790259e-02,\n",
       "         8.56068954e-02,  -2.77291909e-02,   3.21956351e-02,\n",
       "        -1.44229010e-02,   9.81396952e-05,  -1.02728277e-01,\n",
       "         6.77411109e-02,  -8.00238363e-03,  -2.00990103e-02,\n",
       "         6.84855133e-02,   2.01920606e-02,   1.26549322e-02,\n",
       "         4.91309129e-02,   3.32948024e-04,  -3.05207185e-02,\n",
       "        -9.00733396e-02,   3.39636044e-03,   2.28905398e-02,\n",
       "         5.28529510e-02,  -4.48505692e-02,   4.11285311e-02,\n",
       "         7.55573884e-02,   4.83865058e-03,   4.87587117e-02,\n",
       "        -5.28529510e-02,   1.21896779e-02,  -1.00029800e-02,\n",
       "         6.14136420e-02,  -4.76420969e-02,  -5.88082150e-02,\n",
       "        -6.73689023e-02,   4.18729372e-02,  -2.79152915e-02,\n",
       "         1.45159522e-02,   1.55395130e-02,  -6.83924649e-03,\n",
       "        -3.31494084e-04,  -6.17858469e-02,  -2.71708835e-02,\n",
       "         1.34923914e-02,  -8.00238363e-03,   1.28410347e-02,\n",
       "         8.74679163e-02,  -5.91804199e-02,  -4.91309129e-02,\n",
       "        -7.55573884e-02,   1.42181888e-01,  -6.32746592e-02,\n",
       "        -6.66244999e-02,   1.27479834e-02,  -1.40739593e-03,\n",
       "        -4.44783643e-02,  -4.35013277e-03,   7.25797564e-02,\n",
       "         1.99129079e-02,   5.95526211e-03,  -1.38645954e-02,\n",
       "        -5.65749928e-02,   5.28529510e-02,  -1.54464617e-02,\n",
       "        -4.13146317e-02,  -1.48881553e-02,  -1.34737805e-01,\n",
       "         4.37339582e-02,   2.15180381e-03,  -7.48129860e-02,\n",
       "         6.25302568e-02,   3.85231040e-02,   2.88458019e-02,\n",
       "         3.18234339e-02,  -3.38705555e-02,   9.11899507e-02,\n",
       "        -4.28034477e-02,  -6.32746592e-02,  -3.83369997e-02,\n",
       "         6.29024580e-02,   6.40190691e-02,  -4.22451422e-02,\n",
       "         1.10730659e-02,  -1.30271362e-02,  -1.35854417e-02,\n",
       "        -5.65749928e-02,   1.10172354e-01,   8.30014646e-02,\n",
       "        -3.23817395e-02,   7.44407773e-02,  -1.69352777e-02,\n",
       "         4.98753227e-02,   1.60047673e-02,  -1.90754491e-03,\n",
       "        -3.83369997e-02,   2.25183349e-02,   8.63513052e-02,\n",
       "         3.83369997e-02,  -3.72203887e-02,   4.42922637e-02,\n",
       "        -1.19105242e-01,  -6.36468679e-02,   4.80143018e-02,\n",
       "        -2.49376614e-02,  -9.11899507e-02,   9.97506455e-02,\n",
       "        -1.42367985e-02,  -8.11404511e-02,  -5.35973608e-02,\n",
       "        -1.49812065e-02,   8.11404511e-02,  -2.56820694e-02,\n",
       "         2.97763105e-02,  -3.72203887e-02,   2.97763105e-02,\n",
       "         9.04455483e-02,   1.20035755e-02,  -2.15878263e-02,\n",
       "        -8.11404511e-02,   2.10295208e-02,   2.47515589e-02,\n",
       "         2.99624130e-02,   6.73689023e-02,  -7.03465343e-02,\n",
       "         9.90062356e-02,  -1.38645954e-02,   1.04217090e-01,\n",
       "         3.44288610e-02,   2.26113871e-02,  -4.76420969e-02,\n",
       "         1.72144305e-02,   1.74935833e-02,   3.03346179e-02,\n",
       "         2.43793558e-02,   6.99743330e-02,   6.69967011e-02,\n",
       "        -4.91309129e-02,  -8.56068917e-03,  -6.40190691e-02,\n",
       "         9.95645393e-03,   3.29400450e-02,   1.64700225e-02,\n",
       "         1.47951050e-02,   7.85350204e-02,  -4.46644686e-02,\n",
       "        -9.52841938e-02,   7.18353540e-02,  -1.87962968e-02,\n",
       "        -3.40566561e-02,  -9.81687754e-03,   6.17858469e-02,\n",
       "         3.68481837e-02,  -7.66739994e-02,   1.25618810e-02,\n",
       "         1.31201874e-02,   1.02821328e-02,  -2.25183349e-02,\n",
       "        -3.04741948e-03,  -9.81687754e-03,  -6.47634789e-02,\n",
       "         5.28529510e-02,   4.72698957e-02,  -8.46763887e-03], dtype=float32)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### obtain vector for a specific word\n",
    "wv.syn0norm[wv.vocab['unhappy'].index] # (300,)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Other fun operations you can do with word2vec model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "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": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wv.most_similar(positive=['woman', 'king'], negative=['man'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Republican', 0.571267306804657),\n",
       " ('Republicans', 0.5465080738067627),\n",
       " ('Donald_Trump', 0.5382609367370605),\n",
       " ('Gingrich', 0.5238125324249268),\n",
       " ('Palin', 0.5126574635505676),\n",
       " ('Santorum', 0.4977916479110718),\n",
       " ('Dems', 0.49728360772132874),\n",
       " ('Democratic', 0.49602842330932617),\n",
       " ('GOPers', 0.49341505765914917),\n",
       " ('Democrats', 0.4912267029285431)]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wv.most_similar(positive=['GOP', 'Trump'], negative=['money'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Try to research mode about the intuition behind these operations and try to come up with more interesting ones."
   ]
  },
  {
   "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",
    "\n",
    "![alt txt](https://docs.google.com/drawings/d/1QI_5_uxopOYgk2PzU_RMTUGKkoNhJp81aIz4SUwsHIM/pub?w=956&h=263)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.76924075586436913"
      ]
     },
     "execution_count": 59,
     "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": [
    "**Exercise:** Try to find cosine similarity between emotion related words. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### KNN Classification\n",
    "\n",
    "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": 60,
   "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": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### train model\n",
    "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.emotion)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "### obtain predictions on test data\n",
    "predicted = knn_naive_dv.predict(X_test_word_average)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.394812680115\n",
      "confusion matrix\n",
      " [[38 29  9  8]\n",
      " [35 52 12 11]\n",
      " [20 19 29 11]\n",
      " [28 16 12 18]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVcAAAEmCAYAAADWT9N8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nB0muERF9kOQaEdEHSa4REX3w/wPXlh3H6W6GLgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fec15044c88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### evaluate the model\n",
    "word_embeddings.evaluate_prediction(predicted, test_data.emotion, my_tags)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can see that K-Neareset Neighbour doesn't perform well when using the average vector representations (obtained from pretrained word2vec) of the text data. This is partly because the pretrained word2vec may not be helping with the task at hand. What if we train our own word2vec using our dataset? (We need a lot of data!)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"I can't WAIT to go to work tomorrow with a high as fuck fever. [sarcasm]\\\\nHopefully I'll feel better tomorrow.\\\\nBut I doubt it. #pessimism\""
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data.iloc()[56]['text']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('anyway', 0.6451989412307739),\n",
       " ('do', 0.637501060962677),\n",
       " ('just', 0.6333133578300476),\n",
       " ('maybe', 0.6231459379196167),\n",
       " ('guess', 0.6176087856292725),\n",
       " ('it', 0.614241361618042),\n",
       " ('suppose', 0.6105259656906128),\n",
       " ('so', 0.6092208623886108),\n",
       " ('think', 0.6042969822883606),\n",
       " (\"'ll\", 0.5944103002548218),\n",
       " ('want', 0.5808286666870117),\n",
       " ('probably', 0.5804653167724609),\n",
       " ('really', 0.5771951675415039),\n",
       " ('know', 0.576975405216217),\n",
       " ('going', 0.5763067603111267),\n",
       " ('yeah', 0.5684106349945068),\n",
       " ('I', 0.5659971237182617),\n",
       " ('go', 0.5607538223266602),\n",
       " ('even', 0.560643196105957),\n",
       " ('if', 0.5593100190162659)]"
      ]
     },
     "execution_count": 65,
     "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 result is that the average of this particular document falls under some area that is not related to the overall emotion portrayed by the original document."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Try to remove stopwords. Try to figure out how you would achieve this using the current implementation of the *w2v_tokenize_text* function. Also, report the accuracy you achieved."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Head on over to the Gensim website and figure out how to train a wordvec on your own using the gensim library. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "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": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "from gensim.models import Doc2Vec\n",
    "from gensim.models.doc2vec import TaggedDocument"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "### First we tag each text with its emotion\n",
    "train_tagged = train_data.apply(\n",
    "    lambda r: TaggedDocument(words=word_embeddings.tokenize_text(r['text']), tags=[r.emotion]), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_tagged = test_data.apply(\n",
    "    lambda r: TaggedDocument(words=word_embeddings.tokenize_text(r['text']), tags=[r.emotion]), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TaggedDocument(words=['chutneysupercat', 'hi', 'lovely', 'brownie', 'mm', 'is', 'calling', 'me', 'tuppytupperware..', 'its', 'awful'], tags=['fear'])"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_tagged.values[50]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "# transform into proper format\n",
    "trainsent = train_tagged.values\n",
    "testsent = test_tagged.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ TaggedDocument(words=['recording', 'some', 'more', 'fnaf', 'and', 'had', 'to', 'facetime', 'my', 'mum', 'to', 'let', 'her', 'know', 'was', 'okay', 'after', 'let', 'out', 'high', 'pitched', 'scream', 'suchagirl'], tags=['fear']),\n",
       "       TaggedDocument(words=['darwinwatersons', 'pennyfitzger31', 'gumballwatter15', 'ok', 'but', 'why', 'tomorrow', 'horrible', 'day', 'my', 'day', 'was', 'great'], tags=['fear']),\n",
       "       TaggedDocument(words=['budget', 'car', 'rental', 'you', 'have', 'made', 'realize', 'why', 'always', 'use', 'nationalcares', 'what', 'nightmare', 'disgusted'], tags=['fear']),\n",
       "       ...,\n",
       "       TaggedDocument(words=['lennyabrahamson', 'may', 'send', 'you', 'copy', 'of', 'herothegreyhound', 'either', 'e-book', 'or', 'real', 'paper', 'one', 'boy', 'and', 'greyhound', 'smiles', 'tears'], tags=['joy']),\n",
       "       TaggedDocument(words=['regret', 'for', 'the', 'things', 'we', 'did', 'can', 'be', 'tempered', 'by', 'time', 'it', 'is', 'regret', 'for', 'the', 'things', 'we', 'did', 'not', 'do', 'that', 'is', 'inconsolable', 'sydney', 'j.', 'harris'], tags=['sadness']),\n",
       "       TaggedDocument(words=['sami_3499', 'oh', 'noooo', 'nomorehammocks', 'nightmare'], tags=['fear'])], dtype=object)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainsent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-11-19 23:12:39,197 : WARNING : consider setting layer size to a multiple of 4 for greater performance\n",
      "2017-11-19 23:12:39,198 : INFO : collecting all words and their counts\n",
      "2017-11-19 23:12:39,198 : INFO : PROGRESS: at example #0, processed 0 words (0/s), 0 word types, 0 tags\n",
      "2017-11-19 23:12:39,219 : INFO : collected 10719 word types and 4 unique tags from a corpus of 3613 examples and 56393 words\n",
      "2017-11-19 23:12:39,220 : INFO : Loading a fresh vocabulary\n",
      "2017-11-19 23:12:39,225 : INFO : min_count=5 retains 1361 unique words (12% of original 10719, drops 9358)\n",
      "2017-11-19 23:12:39,225 : INFO : min_count=5 leaves 42365 word corpus (75% of original 56393, drops 14028)\n",
      "2017-11-19 23:12:39,228 : INFO : deleting the raw counts dictionary of 10719 items\n",
      "2017-11-19 23:12:39,229 : INFO : sample=0.001 downsamples 63 most-common words\n",
      "2017-11-19 23:12:39,230 : INFO : downsampling leaves estimated 31357 word corpus (74.0% of prior 42365)\n",
      "2017-11-19 23:12:39,230 : INFO : estimated required memory for 1361 words and 5 dimensions: 735820 bytes\n",
      "2017-11-19 23:12:39,233 : INFO : resetting layer weights\n",
      "2017-11-19 23:12:39,246 : INFO : training model with 1 workers on 1361 vocabulary and 5 features, using sg=0 hs=0 sample=0.001 negative=5 window=5\n",
      "2017-11-19 23:12:40,261 : INFO : PROGRESS: at 65.57% examples, 452991 words/s, in_qsize 2, out_qsize 0\n",
      "2017-11-19 23:12:40,791 : INFO : worker thread finished; awaiting finish of 0 more threads\n",
      "2017-11-19 23:12:40,791 : INFO : training on 1127860 raw words (699604 effective words) took 1.5s, 453417 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": 77,
   "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": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('fear', 'anger', 'sadness', 'joy', 'fear')"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_targets[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([-0.39524838, -0.08373364, -0.14480482,  0.27350155,  0.20739011], dtype=float32),\n",
       " array([-0.37708506, -0.2205207 ,  0.30018979,  0.21298207, -0.10291009], dtype=float32),\n",
       " array([-0.10260689,  0.04069313,  0.1139339 ,  0.16602848, -0.02557355], dtype=float32),\n",
       " array([-0.4432691 ,  0.125214  , -0.04396084,  0.25234041,  0.12425233], dtype=float32),\n",
       " array([ 0.09505802,  0.09512238, -0.03159134,  0.17345698,  0.09683908], dtype=float32))"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_regressors[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train a logistic regression model\n",
    "d2v_model = linear_model.LogisticRegression(n_jobs=1, C=1e5)\n",
    "d2v_model = d2v_model.fit(train_regressors, train_targets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.590778097983\n",
      "confusion matrix\n",
      " [[40 25  9 10]\n",
      " [15 77 11  7]\n",
      " [ 6 14 50  9]\n",
      " [10 16 10 38]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAEmCAYAAADbUaM7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nOImamTXASdTMrAFOomZmDfj/ZYI35rTZcEAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fec141ce2e8>"
      ]
     },
     "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."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-11-19 23:15:05,294 : INFO : precomputing L2-norms of doc weight vectors\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.590778097983\n",
      "confusion matrix\n",
      " [[38 18 15 13]\n",
      " [14 72 15  9]\n",
      " [ 6 10 54  9]\n",
      " [ 6 11 16 41]]\n",
      "(row=expected, col=predicted)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAEmCAYAAADbUaM7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nM2vA/we3asq4netq+wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fed0b417eb8>"
      ]
     },
     "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": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('sadness', 0.2729802429676056),\n",
       " ('fear', 0.247198224067688),\n",
       " ('joy', 0.2352927327156067)]"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doc2vec_model.docvecs.most_similar('anger')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('sadness', 0.3019334375858307),\n",
       " ('anger', 0.2471982091665268),\n",
       " ('joy', 0.23824873566627502)]"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doc2vec_model.docvecs.most_similar('fear')"
   ]
  },
  {
   "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 describe the tag."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('strength', 0.9957280158996582),\n",
       " ('gloomy', 0.978492021560669),\n",
       " ('beautiful', 0.9719856977462769),\n",
       " ('present', 0.9529615640640259),\n",
       " ('lol', 0.9487402439117432),\n",
       " ('thenicebot', 0.948267936706543),\n",
       " ('letting', 0.9436185359954834),\n",
       " ('haunt', 0.9382511377334595),\n",
       " ('accept', 0.9362047910690308),\n",
       " ('chirp', 0.9357328414916992)]"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "doc2vec_model.most_similar([doc2vec_model.docvecs['joy']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization\n",
    "We will visualize embeddings using a popular algorithm called T-SNE, which is very similar to the dimentionality reduction algorithm called PCA."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "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": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "### unpacking words from the dco2vec model trained\n",
    "doc2vec_words = word2vec_helpers.unpack_words_from_doc_vector(doc2vec_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3YAAAJCCAYAAACMMVWCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIiDU2AnIiIiIiLi4P4/MFK5CTA4ecMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fec14993358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word2vec_helpers.visualize_vectors(doc2vec_model, doc2vec_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise:** Try to visualize words and find interesting clusters. There are so many creative ways to visualize these powerful vector representations. You can also try other type of classifiers and apply other improvements in the preprocessing stages."
   ]
  },
  {
   "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
}