imdbDataset[SOLVED].ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package words to\n",
      "[nltk_data]     C:\\Users\\Atta\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Unzipping corpora\\words.zip.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import nltk\n",
    "nltk.download('words')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting plotly\n",
      "  Downloading https://files.pythonhosted.org/packages/70/19/8437e22c84083a6d5d8a3c80f4edc73c9dcbb89261d07e6bd13b48752bbd/plotly-4.1.1-py2.py3-none-any.whl (7.1MB)\n",
      "Requirement already satisfied: six in c:\\programdata\\anaconda3\\lib\\site-packages (from plotly) (1.12.0)\n",
      "Collecting retrying>=1.3.3 (from plotly)\n",
      "  Downloading https://files.pythonhosted.org/packages/44/ef/beae4b4ef80902f22e3af073397f079c96969c69b2c7d52a57ea9ae61c9d/retrying-1.3.3.tar.gz\n",
      "Building wheels for collected packages: retrying\n",
      "  Building wheel for retrying (setup.py): started\n",
      "  Building wheel for retrying (setup.py): finished with status 'done'\n",
      "  Stored in directory: C:\\Users\\Atta\\AppData\\Local\\pip\\Cache\\wheels\\d7\\a9\\33\\acc7b709e2a35caa7d4cae442f6fe6fbf2c43f80823d46460c\n",
      "Successfully built retrying\n",
      "Installing collected packages: retrying, plotly\n",
      "Successfully installed plotly-4.1.1 retrying-1.3.3\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install plotly"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "        <script type=\"text/javascript\">\n",
       "        window.PlotlyConfig = {MathJaxConfig: 'local'};\n",
       "        if (window.MathJax) {MathJax.Hub.Config({SVG: {font: \"STIX-Web\"}});}\n",
       "        if (typeof require !== 'undefined') {\n",
       "        require.undef(\"plotly\");\n",
       "        requirejs.config({\n",
       "            paths: {\n",
       "                'plotly': ['https://cdn.plot.ly/plotly-latest.min']\n",
       "            }\n",
       "        });\n",
       "        require(['plotly'], function(Plotly) {\n",
       "            window._Plotly = Plotly;\n",
       "        });\n",
       "        }\n",
       "        </script>\n",
       "        "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import re\n",
    "import os\n",
    "from IPython.display import HTML\n",
    "\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.feature_extraction import text \n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "from tensorflow.python.keras.models import Sequential, load_model\n",
    "from tensorflow.python.keras.layers import Dense, Dropout\n",
    "from tensorflow.python.keras import optimizers\n",
    "\n",
    "import nltk\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "from nltk.stem.snowball import SnowballStemmer\n",
    "from nltk.corpus import words\n",
    "from nltk.corpus import wordnet \n",
    "allEnglishWords = words.words() + [w for w in wordnet.words()]\n",
    "allEnglishWords = np.unique([x.lower() for x in allEnglishWords])\n",
    "\n",
    "import plotly.offline as py\n",
    "import plotly.graph_objs as go\n",
    "py.init_notebook_mode(connected=True)\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "path = \"D:/Deeplearning/datasets/imdb-movie-reviews-dataset/aclImdb/\"\n",
    "positiveFiles = [x for x in os.listdir(path+\"train/pos/\") if x.endswith(\".txt\")]\n",
    "negativeFiles = [x for x in os.listdir(path+\"train/neg/\") if x.endswith(\".txt\")]\n",
    "testFiles = [x for x in os.listdir(path+\"test/\") if x.endswith(\".txt\")]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "positiveReviews, negativeReviews, testReviews = [], [], []\n",
    "for pfile in positiveFiles:\n",
    "    with open(path+\"train/pos/\"+pfile, encoding=\"latin1\") as f:\n",
    "        positiveReviews.append(f.read())\n",
    "for nfile in negativeFiles:\n",
    "    with open(path+\"train/neg/\"+nfile, encoding=\"latin1\") as f:\n",
    "        negativeReviews.append(f.read())\n",
    "for tfile in testFiles:\n",
    "    with open(path+\"test/\"+tfile, encoding=\"latin1\") as f:\n",
    "        testReviews.append(f.read())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21939</th>\n",
       "      <td>Gwyneth Paltrow is absolutely great in this mo...</td>\n",
       "      <td>0</td>\n",
       "      <td>7246_4.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24113</th>\n",
       "      <td>I own this movie. Not by choice, I do. I was r...</td>\n",
       "      <td>0</td>\n",
       "      <td>9202_1.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4633</th>\n",
       "      <td>Well I guess it supposedly not a classic becau...</td>\n",
       "      <td>1</td>\n",
       "      <td>2920_8.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17240</th>\n",
       "      <td>I am, as many are, a fan of Tony Scott films. ...</td>\n",
       "      <td>0</td>\n",
       "      <td>3016_1.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4894</th>\n",
       "      <td>I wish \"that '70s show\" would come back on tel...</td>\n",
       "      <td>1</td>\n",
       "      <td>3155_10.txt</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label         file\n",
       "21939  Gwyneth Paltrow is absolutely great in this mo...      0   7246_4.txt\n",
       "24113  I own this movie. Not by choice, I do. I was r...      0   9202_1.txt\n",
       "4633   Well I guess it supposedly not a classic becau...      1   2920_8.txt\n",
       "17240  I am, as many are, a fan of Tony Scott films. ...      0   3016_1.txt\n",
       "4894   I wish \"that '70s show\" would come back on tel...      1  3155_10.txt"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reviews = pd.concat([\n",
    "    pd.DataFrame({\"review\":positiveReviews, \"label\":1, \"file\":positiveFiles}),\n",
    "    pd.DataFrame({\"review\":negativeReviews, \"label\":0, \"file\":negativeFiles}),\n",
    "    pd.DataFrame({\"review\":testReviews, \"label\":-1, \"file\":testFiles})\n",
    "], ignore_index=True).sample(frac=1, random_state=1)\n",
    "reviews.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "reviews = reviews[[\"review\", \"label\", \"file\"]].sample(frac=1, random_state=1)\n",
    "train = reviews[reviews.label!=-1].sample(frac=0.6, random_state=1)\n",
    "valid = reviews[reviews.label!=-1].drop(train.index)\n",
    "test = reviews[reviews.label==-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(15000, 3)\n",
      "(10000, 3)\n",
      "(2, 3)\n"
     ]
    }
   ],
   "source": [
    "print(train.shape)\n",
    "print(valid.shape)\n",
    "print(test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "One word can describe this movie and that is weird. I recorded this movie one day because it was a Japanese animation and it was old so I thought it would be interesting. Well it was, the movie is about a young boy who travels the universe to get a metal body so he can seek revenge. On the way he meets very colorful characters and must ultimately decide if he wants the body or not. Very strange, if you are a fan of animation/science-fiction you might want to check this out."
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "HTML(train.review.iloc[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Preprocessor(object):\n",
    "    ''' Preprocess data for NLP tasks. '''\n",
    "\n",
    "    def __init__(self, alpha=True, lower=True, stemmer=True, english=False):\n",
    "        self.alpha = alpha\n",
    "        self.lower = lower\n",
    "        self.stemmer = stemmer\n",
    "        self.english = english\n",
    "        \n",
    "        self.uniqueWords = None\n",
    "        self.uniqueStems = None\n",
    "        \n",
    "    def fit(self, texts):\n",
    "        texts = self._doAlways(texts)\n",
    "\n",
    "        allwords = pd.DataFrame({\"word\": np.concatenate(texts.apply(lambda x: x.split()).values)})\n",
    "        self.uniqueWords = allwords.groupby([\"word\"]).size().rename(\"count\").reset_index()\n",
    "        self.uniqueWords = self.uniqueWords[self.uniqueWords[\"count\"]>1]\n",
    "        if self.stemmer:\n",
    "            self.uniqueWords[\"stem\"] = self.uniqueWords.word.apply(lambda x: PorterStemmer().stem(x)).values\n",
    "            self.uniqueWords.sort_values([\"stem\", \"count\"], inplace=True, ascending=False)\n",
    "            self.uniqueStems = self.uniqueWords.groupby(\"stem\").first()\n",
    "        \n",
    "        #if self.english: self.words[\"english\"] = np.in1d(self.words[\"mode\"], allEnglishWords)\n",
    "        print(\"Fitted.\")\n",
    "            \n",
    "    def transform(self, texts):\n",
    "        texts = self._doAlways(texts)\n",
    "        if self.stemmer:\n",
    "            allwords = np.concatenate(texts.apply(lambda x: x.split()).values)\n",
    "            uniqueWords = pd.DataFrame(index=np.unique(allwords))\n",
    "            uniqueWords[\"stem\"] = pd.Series(uniqueWords.index).apply(lambda x: PorterStemmer().stem(x)).values\n",
    "            uniqueWords[\"mode\"] = uniqueWords.stem.apply(lambda x: self.uniqueStems.loc[x, \"word\"] if x in self.uniqueStems.index else \"\")\n",
    "            texts = texts.apply(lambda x: \" \".join([uniqueWords.loc[y, \"mode\"] for y in x.split()]))\n",
    "        #if self.english: texts = self.words.apply(lambda x: \" \".join([y for y in x.split() if self.words.loc[y,\"english\"]]))\n",
    "        print(\"Transformed.\")\n",
    "        return(texts)\n",
    "\n",
    "    def fit_transform(self, texts):\n",
    "        texts = self._doAlways(texts)\n",
    "        self.fit(texts)\n",
    "        texts = self.transform(texts)\n",
    "        return(texts)\n",
    "    \n",
    "    def _doAlways(self, texts):\n",
    "        # Remove parts between <>'s\n",
    "        texts = texts.apply(lambda x: re.sub('<.*?>', ' ', x))\n",
    "        # Keep letters and digits only.\n",
    "        if self.alpha: texts = texts.apply(lambda x: re.sub('[^a-zA-Z0-9 ]+', ' ', x))\n",
    "        # Set everything to lower case\n",
    "        if self.lower: texts = texts.apply(lambda x: x.lower())\n",
    "        return texts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6011</th>\n",
       "      <td>One word can describe this movie and that is w...</td>\n",
       "      <td>1</td>\n",
       "      <td>4160_9.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9653</th>\n",
       "      <td>The Ancient Mariner is a truly classic piece o...</td>\n",
       "      <td>1</td>\n",
       "      <td>7439_9.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15040</th>\n",
       "      <td>The late 80's saw an inexplicable rash of supe...</td>\n",
       "      <td>0</td>\n",
       "      <td>12287_3.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6029</th>\n",
       "      <td>A delightful piece of cinema storytelling in a...</td>\n",
       "      <td>1</td>\n",
       "      <td>4177_9.txt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9729</th>\n",
       "      <td>Greetings again from the darkness. Mary Heron ...</td>\n",
       "      <td>1</td>\n",
       "      <td>7507_8.txt</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label         file\n",
       "6011   One word can describe this movie and that is w...      1   4160_9.txt\n",
       "9653   The Ancient Mariner is a truly classic piece o...      1   7439_9.txt\n",
       "15040  The late 80's saw an inexplicable rash of supe...      0  12287_3.txt\n",
       "6029   A delightful piece of cinema storytelling in a...      1   4177_9.txt\n",
       "9729   Greetings again from the darkness. Mary Heron ...      1   7507_8.txt"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "preprocess = Preprocessor(alpha=True, lower=True, stemmer=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitted.\n",
      "Transformed.\n",
      "Transformed.\n",
      "Wall time: 2min 32s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "trainX = preprocess.fit_transform(train.review)\n",
    "validX = preprocess.transform(valid.review)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6011     one word can describe this movie and that is w...\n",
       "9653     the ancient marine is a truly classic piece of...\n",
       "15040    the late 80 s saw an inexplicable rash of supe...\n",
       "6029     a delightful piece of cinema storytelling in a...\n",
       "9729     greetings again from the dark mary  is amassed...\n",
       "Name: review, dtype: object"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainX.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(38123, 3)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>word</th>\n",
       "      <th>count</th>\n",
       "      <th>stem</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15094</th>\n",
       "      <td>disapproving</td>\n",
       "      <td>6</td>\n",
       "      <td>disapprov</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15093</th>\n",
       "      <td>disapproves</td>\n",
       "      <td>5</td>\n",
       "      <td>disapprov</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15091</th>\n",
       "      <td>disapproval</td>\n",
       "      <td>3</td>\n",
       "      <td>disapprov</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15087</th>\n",
       "      <td>disappointingly</td>\n",
       "      <td>12</td>\n",
       "      <td>disappointingli</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15085</th>\n",
       "      <td>disappointed</td>\n",
       "      <td>569</td>\n",
       "      <td>disappoint</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15086</th>\n",
       "      <td>disappointing</td>\n",
       "      <td>260</td>\n",
       "      <td>disappoint</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15088</th>\n",
       "      <td>disappointment</td>\n",
       "      <td>247</td>\n",
       "      <td>disappoint</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15084</th>\n",
       "      <td>disappoint</td>\n",
       "      <td>58</td>\n",
       "      <td>disappoint</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15090</th>\n",
       "      <td>disappoints</td>\n",
       "      <td>21</td>\n",
       "      <td>disappoint</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15089</th>\n",
       "      <td>disappointments</td>\n",
       "      <td>13</td>\n",
       "      <td>disappoint</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15081</th>\n",
       "      <td>disappeared</td>\n",
       "      <td>61</td>\n",
       "      <td>disappear</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15078</th>\n",
       "      <td>disappear</td>\n",
       "      <td>58</td>\n",
       "      <td>disappear</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15083</th>\n",
       "      <td>disappears</td>\n",
       "      <td>44</td>\n",
       "      <td>disappear</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15079</th>\n",
       "      <td>disappearance</td>\n",
       "      <td>21</td>\n",
       "      <td>disappear</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15082</th>\n",
       "      <td>disappearing</td>\n",
       "      <td>17</td>\n",
       "      <td>disappear</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15080</th>\n",
       "      <td>disappearances</td>\n",
       "      <td>4</td>\n",
       "      <td>disappear</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  word  count             stem\n",
       "15094     disapproving      6        disapprov\n",
       "15093      disapproves      5        disapprov\n",
       "15091      disapproval      3        disapprov\n",
       "15087  disappointingly     12  disappointingli\n",
       "15085     disappointed    569       disappoint\n",
       "15086    disappointing    260       disappoint\n",
       "15088   disappointment    247       disappoint\n",
       "15084       disappoint     58       disappoint\n",
       "15090      disappoints     21       disappoint\n",
       "15089  disappointments     13       disappoint\n",
       "15081      disappeared     61        disappear\n",
       "15078        disappear     58        disappear\n",
       "15083       disappears     44        disappear\n",
       "15079    disappearance     21        disappear\n",
       "15082     disappearing     17        disappear\n",
       "15080   disappearances      4        disappear"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(preprocess.uniqueWords.shape)\n",
    "preprocess.uniqueWords[preprocess.uniqueWords.word.str.contains(\"disapp\")]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(25274, 2)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>word</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>stem</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>disappoint</th>\n",
       "      <td>disappointed</td>\n",
       "      <td>569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>disappointingli</th>\n",
       "      <td>disappointingly</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            word  count\n",
       "stem                                   \n",
       "disappoint          disappointed    569\n",
       "disappointingli  disappointingly     12"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(preprocess.uniqueStems.shape)\n",
    "preprocess.uniqueStems[preprocess.uniqueStems.word.str.contains(\"disappoint\")]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "stop_words = text.ENGLISH_STOP_WORDS.union([\"thats\",\"weve\",\"dont\",\"lets\",\"youre\",\"im\",\"thi\",\"ha\",\n",
    "    \"wa\",\"st\",\"ask\",\"want\",\"like\",\"thank\",\"know\",\"susan\",\"ryan\",\"say\",\"got\",\"ought\",\"ive\",\"theyre\"])\n",
    "tfidf = TfidfVectorizer(min_df=2, max_features=10000, stop_words=stop_words) #, ngram_range=(1,3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 10.9 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "trainX = tfidf.fit_transform(trainX).toarray()\n",
    "validX = tfidf.transform(validX).toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(15000, 10000)\n",
      "(10000, 10000)\n"
     ]
    }
   ],
   "source": [
    "print(trainX.shape)\n",
    "print(validX.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainY = train.label\n",
    "validY = valid.label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(15000, 10000) (15000,)\n",
      "(10000, 10000) (10000,)\n"
     ]
    }
   ],
   "source": [
    "print(trainX.shape, trainY.shape)\n",
    "print(validX.shape, validY.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.stats.stats import pearsonr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[-0.01133404 -0.02084958  0.01552458 ...  0.02185732  0.00800117\n",
      " -0.00226902]\n"
     ]
    }
   ],
   "source": [
    "getCorrelation = np.vectorize(lambda x: pearsonr(trainX[:,x], trainY)[0])\n",
    "correlations = getCorrelation(np.arange(trainX.shape[1]))\n",
    "print(correlations)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "allIndeces = np.argsort(-correlations)\n",
    "bestIndeces = allIndeces[np.concatenate([np.arange(1000), np.arange(-1000, 0)])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['great' 'love' 'excellent' 'best' 'beautiful' 'perfect' 'favorite'\n",
      " 'enjoy' 'amazing' 'performance']\n",
      "['minutes' 'stupid' 'horrible' 'terrible' 'boring' 'worse' 'awful' 'waste'\n",
      " 'worst' 'bad']\n"
     ]
    }
   ],
   "source": [
    "vocabulary = np.array(tfidf.get_feature_names())\n",
    "print(vocabulary[bestIndeces][:10])\n",
    "print(vocabulary[bestIndeces][-10:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainX = trainX[:,bestIndeces]\n",
    "validX = validX[:,bestIndeces]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(15000, 2000) (15000,)\n",
      "(10000, 2000) (10000,)\n"
     ]
    }
   ],
   "source": [
    "print(trainX.shape, trainY.shape)\n",
    "print(validX.shape, validY.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\ops\\init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n"
     ]
    }
   ],
   "source": [
    "DROPOUT = 0.5\n",
    "ACTIVATION = \"tanh\"\n",
    "\n",
    "model = Sequential([    \n",
    "    Dense(int(trainX.shape[1]/2), activation=ACTIVATION, input_dim=trainX.shape[1]),\n",
    "    Dropout(DROPOUT),\n",
    "    Dense(int(trainX.shape[1]/2), activation=ACTIVATION, input_dim=trainX.shape[1]),\n",
    "    Dropout(DROPOUT),\n",
    "    Dense(int(trainX.shape[1]/4), activation=ACTIVATION),\n",
    "    Dropout(DROPOUT),\n",
    "    Dense(100, activation=ACTIVATION),\n",
    "    Dropout(DROPOUT),\n",
    "    Dense(20, activation=ACTIVATION),\n",
    "    Dropout(DROPOUT),\n",
    "    Dense(5, activation=ACTIVATION),\n",
    "    Dropout(DROPOUT),\n",
    "    Dense(1, activation='sigmoid'),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\ops\\nn_impl.py:180: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 1000)              2001000   \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 1000)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1000)              1001000   \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 1000)              0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 500)               500500    \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 100)               50100     \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 20)                2020      \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 20)                0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 5)                 105       \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 5)                 0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 1)                 6         \n",
      "=================================================================\n",
      "Total params: 3,554,731\n",
      "Trainable params: 3,554,731\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer=optimizers.Adam(0.00005), loss='binary_crossentropy', metrics=['accuracy'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "EPOCHS = 30\n",
    "BATCHSIZE = 1500"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 15000 samples, validate on 10000 samples\n",
      "Epoch 1/30\n",
      "15000/15000 [==============================] - 11s 756us/sample - loss: 0.7090 - acc: 0.5003 - val_loss: 0.6843 - val_acc: 0.6322\n",
      "Epoch 2/30\n",
      "15000/15000 [==============================] - 9s 607us/sample - loss: 0.6950 - acc: 0.5312 - val_loss: 0.6733 - val_acc: 0.7375\n",
      "Epoch 3/30\n",
      "15000/15000 [==============================] - 9s 581us/sample - loss: 0.6837 - acc: 0.5579 - val_loss: 0.6611 - val_acc: 0.7797\n",
      "Epoch 4/30\n",
      "15000/15000 [==============================] - 9s 607us/sample - loss: 0.6730 - acc: 0.5743 - val_loss: 0.6469 - val_acc: 0.7977\n",
      "Epoch 5/30\n",
      "15000/15000 [==============================] - 9s 576us/sample - loss: 0.6578 - acc: 0.6095 - val_loss: 0.6287 - val_acc: 0.8085\n",
      "Epoch 6/30\n",
      "15000/15000 [==============================] - 10s 675us/sample - loss: 0.6388 - acc: 0.6328 - val_loss: 0.6056 - val_acc: 0.8156\n",
      "Epoch 7/30\n",
      "15000/15000 [==============================] - 10s 659us/sample - loss: 0.6167 - acc: 0.6695 - val_loss: 0.5772 - val_acc: 0.8223\n",
      "Epoch 8/30\n",
      "15000/15000 [==============================] - 8s 566us/sample - loss: 0.5875 - acc: 0.7076 - val_loss: 0.5442 - val_acc: 0.8280\n",
      "Epoch 9/30\n",
      "15000/15000 [==============================] - 9s 618us/sample - loss: 0.5588 - acc: 0.7429 - val_loss: 0.5089 - val_acc: 0.8332\n",
      "Epoch 10/30\n",
      "15000/15000 [==============================] - 9s 584us/sample - loss: 0.5312 - acc: 0.7697 - val_loss: 0.4748 - val_acc: 0.8389\n",
      "Epoch 11/30\n",
      "15000/15000 [==============================] - 9s 574us/sample - loss: 0.5012 - acc: 0.7859 - val_loss: 0.4445 - val_acc: 0.8448\n",
      "Epoch 12/30\n",
      "15000/15000 [==============================] - 9s 584us/sample - loss: 0.4837 - acc: 0.8013 - val_loss: 0.4190 - val_acc: 0.8482\n",
      "Epoch 13/30\n",
      "15000/15000 [==============================] - 9s 581us/sample - loss: 0.4549 - acc: 0.8211 - val_loss: 0.3989 - val_acc: 0.8515\n",
      "Epoch 14/30\n",
      "15000/15000 [==============================] - 9s 592us/sample - loss: 0.4414 - acc: 0.8275 - val_loss: 0.3823 - val_acc: 0.8559\n",
      "Epoch 15/30\n",
      "15000/15000 [==============================] - 9s 575us/sample - loss: 0.4256 - acc: 0.8392 - val_loss: 0.3692 - val_acc: 0.8603\n",
      "Epoch 16/30\n",
      "15000/15000 [==============================] - 9s 620us/sample - loss: 0.4164 - acc: 0.8439 - val_loss: 0.3599 - val_acc: 0.8632\n",
      "Epoch 17/30\n",
      "15000/15000 [==============================] - 9s 597us/sample - loss: 0.4052 - acc: 0.8485 - val_loss: 0.3505 - val_acc: 0.8664\n",
      "Epoch 18/30\n",
      "15000/15000 [==============================] - 9s 590us/sample - loss: 0.3931 - acc: 0.8538 - val_loss: 0.3440 - val_acc: 0.8706\n",
      "Epoch 19/30\n",
      "15000/15000 [==============================] - 9s 574us/sample - loss: 0.3896 - acc: 0.8596 - val_loss: 0.3386 - val_acc: 0.8721\n",
      "Epoch 20/30\n",
      "15000/15000 [==============================] - 9s 580us/sample - loss: 0.3761 - acc: 0.8668 - val_loss: 0.3345 - val_acc: 0.8740\n",
      "Epoch 21/30\n",
      "15000/15000 [==============================] - 9s 591us/sample - loss: 0.3730 - acc: 0.8703 - val_loss: 0.3311 - val_acc: 0.8748\n",
      "Epoch 22/30\n",
      "15000/15000 [==============================] - 9s 597us/sample - loss: 0.3688 - acc: 0.8709 - val_loss: 0.3282 - val_acc: 0.8763\n",
      "Epoch 23/30\n",
      "15000/15000 [==============================] - 9s 579us/sample - loss: 0.3622 - acc: 0.8743 - val_loss: 0.3264 - val_acc: 0.8763\n",
      "Epoch 24/30\n",
      "15000/15000 [==============================] - 9s 579us/sample - loss: 0.3594 - acc: 0.8773 - val_loss: 0.3240 - val_acc: 0.8774\n",
      "Epoch 25/30\n",
      "15000/15000 [==============================] - 9s 569us/sample - loss: 0.3556 - acc: 0.8768 - val_loss: 0.3222 - val_acc: 0.8781\n",
      "Epoch 26/30\n",
      "15000/15000 [==============================] - 9s 592us/sample - loss: 0.3507 - acc: 0.8797 - val_loss: 0.3211 - val_acc: 0.8779\n",
      "Epoch 27/30\n",
      "15000/15000 [==============================] - 9s 588us/sample - loss: 0.3485 - acc: 0.8834 - val_loss: 0.3202 - val_acc: 0.8769\n",
      "Epoch 28/30\n",
      "15000/15000 [==============================] - 8s 564us/sample - loss: 0.3475 - acc: 0.8803 - val_loss: 0.3197 - val_acc: 0.8777\n",
      "Epoch 29/30\n",
      "15000/15000 [==============================] - 9s 614us/sample - loss: 0.3480 - acc: 0.8795 - val_loss: 0.3191 - val_acc: 0.8768\n",
      "Epoch 30/30\n",
      "15000/15000 [==============================] - 9s 597us/sample - loss: 0.3415 - acc: 0.8870 - val_loss: 0.3188 - val_acc: 0.8767\n",
      "Wall time: 4min 32s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x1e71f6350f0>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "model.fit(trainX, trainY, epochs=EPOCHS, batch_size=BATCHSIZE, validation_data=(validX, validY))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "linkText": "Export to plot.ly",
        "plotlyServerURL": "https://plot.ly",
        "showLink": false
       },
       "data": [
        {
         "marker": {
          "size": 5
         },
         "name": "Train Accuracy",
         "type": "scatter",
         "x": [
          0,
          1,
          2,
          3,
          4,
          5,
          6,
          7,
          8,
          9,
          10,
          11,
          12,
          13,
          14,
          15,
          16,
          17,
          18,
          19,
          20,
          21,
          22,
          23,
          24,
          25,
          26,
          27,
          28,
          29
         ],
         "y": [
          0.500333309173584,
          0.5311999917030334,
          0.5579333305358887,
          0.5742666721343994,
          0.6095333099365234,
          0.6327999830245972,
          0.6694666743278503,
          0.7075999975204468,
          0.7429333329200745,
          0.7696666717529297,
          0.7859333157539368,
          0.8012666702270508,
          0.821066677570343,
          0.8274666666984558,
          0.8392000198364258,
          0.8439333438873291,
          0.848466694355011,
          0.8537999987602234,
          0.8596000075340271,
          0.8668000102043152,
          0.8702666759490967,
          0.8708666563034058,
          0.8742666840553284,
          0.8773333430290222,
          0.876800000667572,
          0.8797333240509033,
          0.883400022983551,
          0.8803333044052124,
          0.8795333504676819,
          0.8870000243186951
         ],
         "yaxis": "y2"
        },
        {
         "marker": {
          "size": 5
         },
         "name": "Valid Accuracy",
         "type": "scatter",
         "x": [
          0,
          1,
          2,
          3,
          4,
          5,
          6,
          7,
          8,
          9,
          10,
          11,
          12,
          13,
          14,
          15,
          16,
          17,
          18,
          19,
          20,
          21,
          22,
          23,
          24,
          25,
          26,
          27,
          28,
          29
         ],
         "y": [
          0.6322000026702881,
          0.737500011920929,
          0.779699981212616,
          0.7976999878883362,
          0.8084999918937683,
          0.8155999779701233,
          0.8223000168800354,
          0.828000009059906,
          0.8331999778747559,
          0.8389000296592712,
          0.8447999954223633,
          0.8482000231742859,
          0.8514999747276306,
          0.85589998960495,
          0.8603000044822693,
          0.8632000088691711,
          0.8664000034332275,
          0.8705999851226807,
          0.8720999956130981,
          0.8740000128746033,
          0.8748000264167786,
          0.8762999773025513,
          0.8762999773025513,
          0.8773999810218811,
          0.8780999779701233,
          0.8779000043869019,
          0.8769000172615051,
          0.8776999711990356,
          0.876800000667572,
          0.8766999840736389
         ],
         "yaxis": "y2"
        },
        {
         "marker": {
          "size": 5
         },
         "name": "Train Loss",
         "type": "scatter",
         "x": [
          0,
          1,
          2,
          3,
          4,
          5,
          6,
          7,
          8,
          9,
          10,
          11,
          12,
          13,
          14,
          15,
          16,
          17,
          18,
          19,
          20,
          21,
          22,
          23,
          24,
          25,
          26,
          27,
          28,
          29
         ],
         "y": [
          0.7090335488319397,
          0.6949646353721619,
          0.6837466597557068,
          0.6729557573795318,
          0.6578013002872467,
          0.6388334512710572,
          0.6166866362094879,
          0.5875231266021729,
          0.5587551116943359,
          0.5311960399150848,
          0.5012215822935104,
          0.48370979726314545,
          0.4549483895301819,
          0.4414473086595535,
          0.42558732330799104,
          0.4164389163255692,
          0.40524367690086366,
          0.3930757254362106,
          0.38958784341812136,
          0.3761449843645096,
          0.3729754209518433,
          0.3687675803899765,
          0.3622445225715637,
          0.3593663990497589,
          0.35557909607887267,
          0.3507289856672287,
          0.348482158780098,
          0.3475001692771912,
          0.3480459302663803,
          0.3414720892906189
         ]
        },
        {
         "marker": {
          "size": 5
         },
         "name": "Valid Loss",
         "type": "scatter",
         "x": [
          0,
          1,
          2,
          3,
          4,
          5,
          6,
          7,
          8,
          9,
          10,
          11,
          12,
          13,
          14,
          15,
          16,
          17,
          18,
          19,
          20,
          21,
          22,
          23,
          24,
          25,
          26,
          27,
          28,
          29
         ],
         "y": [
          0.6843045204877853,
          0.6732610464096069,
          0.6611141175031662,
          0.6468547612428666,
          0.6287137240171432,
          0.6056130498647689,
          0.5772046566009521,
          0.5442442685365677,
          0.5088827013969421,
          0.47480556517839434,
          0.4445192217826843,
          0.41897249072790144,
          0.39893068075180055,
          0.3822665438055992,
          0.36920246183872224,
          0.3598753035068512,
          0.3505427807569504,
          0.3439976155757904,
          0.3386452004313469,
          0.3345375508069992,
          0.33111662119627,
          0.3282011434435844,
          0.3264169290661812,
          0.32400280982255936,
          0.3221524551510811,
          0.32105368822813035,
          0.3202482372522354,
          0.3196790710091591,
          0.3191257014870644,
          0.3188201546669006
         ]
        }
       ],
       "layout": {
        "font": {
         "family": "Palatino"
        },
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "scatter": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        },
        "title": {
         "text": "Model Training Evolution"
        },
        "xaxis": {
         "dtick": 1,
         "title": {
          "text": "Epoch"
         }
        },
        "yaxis": {
         "domain": [
          0,
          0.45
         ],
         "title": {
          "text": "Loss"
         }
        },
        "yaxis2": {
         "domain": [
          0.55,
          1
         ],
         "title": {
          "text": "Accuracy"
         }
        }
       }
      },
      "text/html": [
       "<div>\n",
       "        \n",
       "        \n",
       "            <div id=\"0410927b-b380-4bdb-ae6f-f29e6cf100f7\" class=\"plotly-graph-div\" style=\"height:525px; width:100%;\"></div>\n",
       "            <script type=\"text/javascript\">\n",
       "                require([\"plotly\"], function(Plotly) {\n",
       "                    window.PLOTLYENV=window.PLOTLYENV || {};\n",
       "                    \n",
       "                if (document.getElementById(\"0410927b-b380-4bdb-ae6f-f29e6cf100f7\")) {\n",
       "                    Plotly.newPlot(\n",
       "                        '0410927b-b380-4bdb-ae6f-f29e6cf100f7',\n",
       "                        [{\"marker\": {\"size\": 5}, \"name\": \"Train Accuracy\", \"type\": \"scatter\", \"x\": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29], \"y\": [0.500333309173584, 0.5311999917030334, 0.5579333305358887, 0.5742666721343994, 0.6095333099365234, 0.6327999830245972, 0.6694666743278503, 0.7075999975204468, 0.7429333329200745, 0.7696666717529297, 0.7859333157539368, 0.8012666702270508, 0.821066677570343, 0.8274666666984558, 0.8392000198364258, 0.8439333438873291, 0.848466694355011, 0.8537999987602234, 0.8596000075340271, 0.8668000102043152, 0.8702666759490967, 0.8708666563034058, 0.8742666840553284, 0.8773333430290222, 0.876800000667572, 0.8797333240509033, 0.883400022983551, 0.8803333044052124, 0.8795333504676819, 0.8870000243186951], \"yaxis\": \"y2\"}, {\"marker\": {\"size\": 5}, \"name\": \"Valid Accuracy\", \"type\": \"scatter\", \"x\": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29], \"y\": [0.6322000026702881, 0.737500011920929, 0.779699981212616, 0.7976999878883362, 0.8084999918937683, 0.8155999779701233, 0.8223000168800354, 0.828000009059906, 0.8331999778747559, 0.8389000296592712, 0.8447999954223633, 0.8482000231742859, 0.8514999747276306, 0.85589998960495, 0.8603000044822693, 0.8632000088691711, 0.8664000034332275, 0.8705999851226807, 0.8720999956130981, 0.8740000128746033, 0.8748000264167786, 0.8762999773025513, 0.8762999773025513, 0.8773999810218811, 0.8780999779701233, 0.8779000043869019, 0.8769000172615051, 0.8776999711990356, 0.876800000667572, 0.8766999840736389], \"yaxis\": \"y2\"}, {\"marker\": {\"size\": 5}, \"name\": \"Train Loss\", \"type\": \"scatter\", \"x\": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29], \"y\": [0.7090335488319397, 0.6949646353721619, 0.6837466597557068, 0.6729557573795318, 0.6578013002872467, 0.6388334512710572, 0.6166866362094879, 0.5875231266021729, 0.5587551116943359, 0.5311960399150848, 0.5012215822935104, 0.48370979726314545, 0.4549483895301819, 0.4414473086595535, 0.42558732330799104, 0.4164389163255692, 0.40524367690086366, 0.3930757254362106, 0.38958784341812136, 0.3761449843645096, 0.3729754209518433, 0.3687675803899765, 0.3622445225715637, 0.3593663990497589, 0.35557909607887267, 0.3507289856672287, 0.348482158780098, 0.3475001692771912, 0.3480459302663803, 0.3414720892906189]}, {\"marker\": {\"size\": 5}, \"name\": \"Valid Loss\", \"type\": \"scatter\", \"x\": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29], \"y\": [0.6843045204877853, 0.6732610464096069, 0.6611141175031662, 0.6468547612428666, 0.6287137240171432, 0.6056130498647689, 0.5772046566009521, 0.5442442685365677, 0.5088827013969421, 0.47480556517839434, 0.4445192217826843, 0.41897249072790144, 0.39893068075180055, 0.3822665438055992, 0.36920246183872224, 0.3598753035068512, 0.3505427807569504, 0.3439976155757904, 0.3386452004313469, 0.3345375508069992, 0.33111662119627, 0.3282011434435844, 0.3264169290661812, 0.32400280982255936, 0.3221524551510811, 0.32105368822813035, 0.3202482372522354, 0.3196790710091591, 0.3191257014870644, 0.3188201546669006]}],\n",
       "                        {\"font\": {\"family\": \"Palatino\"}, \"template\": {\"data\": {\"bar\": [{\"error_x\": {\"color\": \"#2a3f5f\"}, \"error_y\": {\"color\": \"#2a3f5f\"}, \"marker\": {\"line\": {\"color\": \"#E5ECF6\", \"width\": 0.5}}, \"type\": \"bar\"}], \"barpolar\": [{\"marker\": {\"line\": {\"color\": \"#E5ECF6\", \"width\": 0.5}}, \"type\": \"barpolar\"}], \"carpet\": [{\"aaxis\": {\"endlinecolor\": \"#2a3f5f\", \"gridcolor\": \"white\", \"linecolor\": \"white\", \"minorgridcolor\": \"white\", \"startlinecolor\": \"#2a3f5f\"}, \"baxis\": {\"endlinecolor\": \"#2a3f5f\", \"gridcolor\": \"white\", \"linecolor\": \"white\", \"minorgridcolor\": \"white\", \"startlinecolor\": \"#2a3f5f\"}, \"type\": \"carpet\"}], \"choropleth\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"type\": \"choropleth\"}], \"contour\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"colorscale\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]], \"type\": \"contour\"}], \"contourcarpet\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"type\": \"contourcarpet\"}], \"heatmap\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"colorscale\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]], \"type\": \"heatmap\"}], \"heatmapgl\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"colorscale\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]], \"type\": \"heatmapgl\"}], \"histogram\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"histogram\"}], \"histogram2d\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"colorscale\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]], \"type\": \"histogram2d\"}], \"histogram2dcontour\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"colorscale\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]], \"type\": \"histogram2dcontour\"}], \"mesh3d\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"type\": \"mesh3d\"}], \"parcoords\": [{\"line\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"parcoords\"}], \"scatter\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scatter\"}], \"scatter3d\": [{\"line\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scatter3d\"}], \"scattercarpet\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scattercarpet\"}], \"scattergeo\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scattergeo\"}], \"scattergl\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scattergl\"}], \"scattermapbox\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scattermapbox\"}], \"scatterpolar\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scatterpolar\"}], \"scatterpolargl\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scatterpolargl\"}], \"scatterternary\": [{\"marker\": {\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}}, \"type\": \"scatterternary\"}], \"surface\": [{\"colorbar\": {\"outlinewidth\": 0, \"ticks\": \"\"}, \"colorscale\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]], \"type\": \"surface\"}], \"table\": [{\"cells\": {\"fill\": {\"color\": \"#EBF0F8\"}, \"line\": {\"color\": \"white\"}}, \"header\": {\"fill\": {\"color\": \"#C8D4E3\"}, \"line\": {\"color\": \"white\"}}, \"type\": \"table\"}]}, \"layout\": {\"annotationdefaults\": {\"arrowcolor\": \"#2a3f5f\", \"arrowhead\": 0, \"arrowwidth\": 1}, \"colorscale\": {\"diverging\": [[0, \"#8e0152\"], [0.1, \"#c51b7d\"], [0.2, \"#de77ae\"], [0.3, \"#f1b6da\"], [0.4, \"#fde0ef\"], [0.5, \"#f7f7f7\"], [0.6, \"#e6f5d0\"], [0.7, \"#b8e186\"], [0.8, \"#7fbc41\"], [0.9, \"#4d9221\"], [1, \"#276419\"]], \"sequential\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]], \"sequentialminus\": [[0.0, \"#0d0887\"], [0.1111111111111111, \"#46039f\"], [0.2222222222222222, \"#7201a8\"], [0.3333333333333333, \"#9c179e\"], [0.4444444444444444, \"#bd3786\"], [0.5555555555555556, \"#d8576b\"], [0.6666666666666666, \"#ed7953\"], [0.7777777777777778, \"#fb9f3a\"], [0.8888888888888888, \"#fdca26\"], [1.0, \"#f0f921\"]]}, \"colorway\": [\"#636efa\", \"#EF553B\", \"#00cc96\", \"#ab63fa\", \"#FFA15A\", \"#19d3f3\", \"#FF6692\", \"#B6E880\", \"#FF97FF\", \"#FECB52\"], \"font\": {\"color\": \"#2a3f5f\"}, \"geo\": {\"bgcolor\": \"white\", \"lakecolor\": \"white\", \"landcolor\": \"#E5ECF6\", \"showlakes\": true, \"showland\": true, \"subunitcolor\": \"white\"}, \"hoverlabel\": {\"align\": \"left\"}, \"hovermode\": \"closest\", \"mapbox\": {\"style\": \"light\"}, \"paper_bgcolor\": \"white\", \"plot_bgcolor\": \"#E5ECF6\", \"polar\": {\"angularaxis\": {\"gridcolor\": \"white\", \"linecolor\": \"white\", \"ticks\": \"\"}, \"bgcolor\": \"#E5ECF6\", \"radialaxis\": {\"gridcolor\": \"white\", \"linecolor\": \"white\", \"ticks\": \"\"}}, \"scene\": {\"xaxis\": {\"backgroundcolor\": \"#E5ECF6\", \"gridcolor\": \"white\", \"gridwidth\": 2, \"linecolor\": \"white\", \"showbackground\": true, \"ticks\": \"\", \"zerolinecolor\": \"white\"}, \"yaxis\": {\"backgroundcolor\": \"#E5ECF6\", \"gridcolor\": \"white\", \"gridwidth\": 2, \"linecolor\": \"white\", \"showbackground\": true, \"ticks\": \"\", \"zerolinecolor\": \"white\"}, \"zaxis\": {\"backgroundcolor\": \"#E5ECF6\", \"gridcolor\": \"white\", \"gridwidth\": 2, \"linecolor\": \"white\", \"showbackground\": true, \"ticks\": \"\", \"zerolinecolor\": \"white\"}}, \"shapedefaults\": {\"line\": {\"color\": \"#2a3f5f\"}}, \"ternary\": {\"aaxis\": {\"gridcolor\": \"white\", \"linecolor\": \"white\", \"ticks\": \"\"}, \"baxis\": {\"gridcolor\": \"white\", \"linecolor\": \"white\", \"ticks\": \"\"}, \"bgcolor\": \"#E5ECF6\", \"caxis\": {\"gridcolor\": \"white\", \"linecolor\": \"white\", \"ticks\": \"\"}}, \"title\": {\"x\": 0.05}, \"xaxis\": {\"automargin\": true, \"gridcolor\": \"white\", \"linecolor\": \"white\", \"ticks\": \"\", \"zerolinecolor\": \"white\", \"zerolinewidth\": 2}, \"yaxis\": {\"automargin\": true, \"gridcolor\": \"white\", \"linecolor\": \"white\", \"ticks\": \"\", \"zerolinecolor\": \"white\", \"zerolinewidth\": 2}}}, \"title\": {\"text\": \"Model Training Evolution\"}, \"xaxis\": {\"dtick\": 1, \"title\": {\"text\": \"Epoch\"}}, \"yaxis\": {\"domain\": [0, 0.45], \"title\": {\"text\": \"Loss\"}}, \"yaxis2\": {\"domain\": [0.55, 1], \"title\": {\"text\": \"Accuracy\"}}},\n",
       "                        {\"responsive\": true}\n",
       "                    ).then(function(){\n",
       "                            \n",
       "var gd = document.getElementById('0410927b-b380-4bdb-ae6f-f29e6cf100f7');\n",
       "var x = new MutationObserver(function (mutations, observer) {{\n",
       "        var display = window.getComputedStyle(gd).display;\n",
       "        if (!display || display === 'none') {{\n",
       "            console.log([gd, 'removed!']);\n",
       "            Plotly.purge(gd);\n",
       "            observer.disconnect();\n",
       "        }}\n",
       "}});\n",
       "\n",
       "// Listen for the removal of the full notebook cells\n",
       "var notebookContainer = gd.closest('#notebook-container');\n",
       "if (notebookContainer) {{\n",
       "    x.observe(notebookContainer, {childList: true});\n",
       "}}\n",
       "\n",
       "// Listen for the clearing of the current output cell\n",
       "var outputEl = gd.closest('.output');\n",
       "if (outputEl) {{\n",
       "    x.observe(outputEl, {childList: true});\n",
       "}}\n",
       "\n",
       "                        })\n",
       "                };\n",
       "                });\n",
       "            </script>\n",
       "        </div>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = np.arange(EPOCHS)\n",
    "history = model.history.history\n",
    "\n",
    "data = [\n",
    "    go.Scatter(x=x, y=history[\"acc\"], name=\"Train Accuracy\", marker=dict(size=5), yaxis='y2'),\n",
    "    go.Scatter(x=x, y=history[\"val_acc\"], name=\"Valid Accuracy\", marker=dict(size=5), yaxis='y2'),\n",
    "    go.Scatter(x=x, y=history[\"loss\"], name=\"Train Loss\", marker=dict(size=5)),\n",
    "    go.Scatter(x=x, y=history[\"val_loss\"], name=\"Valid Loss\", marker=dict(size=5))\n",
    "]\n",
    "layout = go.Layout(\n",
    "    title=\"Model Training Evolution\", font=dict(family='Palatino'), xaxis=dict(title='Epoch', dtick=1),\n",
    "    yaxis1=dict(title=\"Loss\", domain=[0, 0.45]), yaxis2=dict(title=\"Accuracy\", domain=[0.55, 1]),\n",
    ")\n",
    "py.iplot(go.Figure(data=data, layout=layout), show_link=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "      <th>probability</th>\n",
       "      <th>prediction</th>\n",
       "      <th>truth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2455</th>\n",
       "      <td>Origins of the Care Bears &amp; their Cousins. If ...</td>\n",
       "      <td>1</td>\n",
       "      <td>1220_9.txt</td>\n",
       "      <td>0.907206</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1043</th>\n",
       "      <td>I always enjoy this movie when it shows up on ...</td>\n",
       "      <td>1</td>\n",
       "      <td>1093_8.txt</td>\n",
       "      <td>0.906141</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17166</th>\n",
       "      <td>I watched this movie when Joe Bob Briggs hoste...</td>\n",
       "      <td>0</td>\n",
       "      <td>2950_1.txt</td>\n",
       "      <td>0.112409</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24532</th>\n",
       "      <td>Former brat pack actor and all round pretty bo...</td>\n",
       "      <td>0</td>\n",
       "      <td>9580_3.txt</td>\n",
       "      <td>0.101270</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23170</th>\n",
       "      <td>Any movie in which Brooke Shields out-acts a F...</td>\n",
       "      <td>0</td>\n",
       "      <td>8354_2.txt</td>\n",
       "      <td>0.095695</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label        file  \\\n",
       "2455   Origins of the Care Bears & their Cousins. If ...      1  1220_9.txt   \n",
       "1043   I always enjoy this movie when it shows up on ...      1  1093_8.txt   \n",
       "17166  I watched this movie when Joe Bob Briggs hoste...      0  2950_1.txt   \n",
       "24532  Former brat pack actor and all round pretty bo...      0  9580_3.txt   \n",
       "23170  Any movie in which Brooke Shields out-acts a F...      0  8354_2.txt   \n",
       "\n",
       "       probability  prediction  truth  \n",
       "2455      0.907206        True   True  \n",
       "1043      0.906141        True   True  \n",
       "17166     0.112409       False  False  \n",
       "24532     0.101270       False  False  \n",
       "23170     0.095695       False  False  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train[\"probability\"] = model.predict(trainX)\n",
    "train[\"prediction\"] = train.probability-0.5>0\n",
    "train[\"truth\"] = train.label==1\n",
    "train.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "15000/15000 [==============================] - 5s 329us/sample - loss: 0.2515 - acc: 0.9220\n",
      "[0.2514871084690094, 0.922]\n",
      "0.922\n"
     ]
    }
   ],
   "source": [
    "print(model.evaluate(trainX, trainY))\n",
    "print((train.truth==train.prediction).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "      <th>probability</th>\n",
       "      <th>prediction</th>\n",
       "      <th>truth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8087</th>\n",
       "      <td>There aren't too many times when I see a film ...</td>\n",
       "      <td>1</td>\n",
       "      <td>6029_7.txt</td>\n",
       "      <td>0.211895</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11702</th>\n",
       "      <td>In Iran women are prohibited from attending li...</td>\n",
       "      <td>1</td>\n",
       "      <td>9283_8.txt</td>\n",
       "      <td>0.889475</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1056</th>\n",
       "      <td>Big S isn't playing with taboos or forcing an ...</td>\n",
       "      <td>1</td>\n",
       "      <td>10951_8.txt</td>\n",
       "      <td>0.894298</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13081</th>\n",
       "      <td>For the first forty minutes, Empire really sha...</td>\n",
       "      <td>0</td>\n",
       "      <td>10523_2.txt</td>\n",
       "      <td>0.096872</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3811</th>\n",
       "      <td>Louise Brooks gives a wonderful performance in...</td>\n",
       "      <td>1</td>\n",
       "      <td>2180_8.txt</td>\n",
       "      <td>0.905861</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label         file  \\\n",
       "8087   There aren't too many times when I see a film ...      1   6029_7.txt   \n",
       "11702  In Iran women are prohibited from attending li...      1   9283_8.txt   \n",
       "1056   Big S isn't playing with taboos or forcing an ...      1  10951_8.txt   \n",
       "13081  For the first forty minutes, Empire really sha...      0  10523_2.txt   \n",
       "3811   Louise Brooks gives a wonderful performance in...      1   2180_8.txt   \n",
       "\n",
       "       probability  prediction  truth  \n",
       "8087      0.211895       False   True  \n",
       "11702     0.889475        True   True  \n",
       "1056      0.894298        True   True  \n",
       "13081     0.096872       False  False  \n",
       "3811      0.905861        True   True  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid[\"probability\"] = model.predict(validX)\n",
    "valid[\"prediction\"] = valid.probability-0.5>0\n",
    "valid[\"truth\"] = valid.label==1\n",
    "valid.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 3s 325us/sample - loss: 0.3188 - acc: 0.8767\n",
      "[0.3188201421737671, 0.8767]\n",
      "0.8767\n"
     ]
    }
   ],
   "source": [
    "print(model.evaluate(validX, validY))\n",
    "print((valid.truth==valid.prediction).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>truth</th>\n",
       "      <th>False</th>\n",
       "      <th>True</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>prediction</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>6919</td>\n",
       "      <td>559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>611</td>\n",
       "      <td>6911</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "truth       False  True \n",
       "prediction              \n",
       "False        6919    559\n",
       "True          611   6911"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainCross = train.groupby([\"prediction\", \"truth\"]).size().unstack()\n",
    "trainCross"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>truth</th>\n",
       "      <th>False</th>\n",
       "      <th>True</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>prediction</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>4339</td>\n",
       "      <td>602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>631</td>\n",
       "      <td>4428</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "truth       False  True \n",
       "prediction              \n",
       "False        4339    602\n",
       "True          631   4428"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validCross = valid.groupby([\"prediction\", \"truth\"]).size().unstack()\n",
    "validCross"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4428 true positives.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "      <th>probability</th>\n",
       "      <th>prediction</th>\n",
       "      <th>truth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>917</th>\n",
       "      <td>I have rarely emerged from viewing a film with...</td>\n",
       "      <td>1</td>\n",
       "      <td>10826_10.txt</td>\n",
       "      <td>0.908975</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2339</th>\n",
       "      <td>Antonio Margheriti's \"Danza Macabra\" aka. \"Cas...</td>\n",
       "      <td>1</td>\n",
       "      <td>12105_10.txt</td>\n",
       "      <td>0.908895</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11228</th>\n",
       "      <td>Director Sidney Lumet has made some masterpiec...</td>\n",
       "      <td>1</td>\n",
       "      <td>8857_10.txt</td>\n",
       "      <td>0.908892</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label          file  \\\n",
       "917    I have rarely emerged from viewing a film with...      1  10826_10.txt   \n",
       "2339   Antonio Margheriti's \"Danza Macabra\" aka. \"Cas...      1  12105_10.txt   \n",
       "11228  Director Sidney Lumet has made some masterpiec...      1   8857_10.txt   \n",
       "\n",
       "       probability  prediction  truth  \n",
       "917       0.908975        True   True  \n",
       "2339      0.908895        True   True  \n",
       "11228     0.908892        True   True  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "truepositives = valid[(valid.truth==True)&(valid.truth==valid.prediction)]\n",
    "print(len(truepositives), \"true positives.\")\n",
    "truepositives.sort_values(\"probability\", ascending=False).head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4339 true negatives.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "      <th>probability</th>\n",
       "      <th>prediction</th>\n",
       "      <th>truth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19935</th>\n",
       "      <td>Aside from the horrendous acting and the ridic...</td>\n",
       "      <td>0</td>\n",
       "      <td>5442_3.txt</td>\n",
       "      <td>0.090714</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20213</th>\n",
       "      <td>I can't believe the positive reviews of this m...</td>\n",
       "      <td>0</td>\n",
       "      <td>5693_1.txt</td>\n",
       "      <td>0.090723</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23196</th>\n",
       "      <td>This movie is most possibly the worst movie I ...</td>\n",
       "      <td>0</td>\n",
       "      <td>8378_1.txt</td>\n",
       "      <td>0.090728</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label        file  \\\n",
       "19935  Aside from the horrendous acting and the ridic...      0  5442_3.txt   \n",
       "20213  I can't believe the positive reviews of this m...      0  5693_1.txt   \n",
       "23196  This movie is most possibly the worst movie I ...      0  8378_1.txt   \n",
       "\n",
       "       probability  prediction  truth  \n",
       "19935     0.090714       False  False  \n",
       "20213     0.090723       False  False  \n",
       "23196     0.090728       False  False  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "truenegatives = valid[(valid.truth==False)&(valid.truth==valid.prediction)]\n",
    "print(len(truenegatives), \"true negatives.\")\n",
    "truenegatives.sort_values(\"probability\", ascending=True).head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "602 false positives.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "      <th>probability</th>\n",
       "      <th>prediction</th>\n",
       "      <th>truth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12340</th>\n",
       "      <td>I wouldn't go so far as to not recommend this ...</td>\n",
       "      <td>1</td>\n",
       "      <td>9858_7.txt</td>\n",
       "      <td>0.091938</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6038</th>\n",
       "      <td>This movie has everything that makes a bad mov...</td>\n",
       "      <td>1</td>\n",
       "      <td>4185_8.txt</td>\n",
       "      <td>0.092445</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7982</th>\n",
       "      <td>After seeing the terrible, terrible, terrible ...</td>\n",
       "      <td>1</td>\n",
       "      <td>5935_9.txt</td>\n",
       "      <td>0.093050</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label        file  \\\n",
       "12340  I wouldn't go so far as to not recommend this ...      1  9858_7.txt   \n",
       "6038   This movie has everything that makes a bad mov...      1  4185_8.txt   \n",
       "7982   After seeing the terrible, terrible, terrible ...      1  5935_9.txt   \n",
       "\n",
       "       probability  prediction  truth  \n",
       "12340     0.091938       False   True  \n",
       "6038      0.092445       False   True  \n",
       "7982      0.093050       False   True  "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "falsepositives = valid[(valid.truth==True)&(valid.truth!=valid.prediction)]\n",
    "print(len(falsepositives), \"false positives.\")\n",
    "falsepositives.sort_values(\"probability\", ascending=True).head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "631 false negatives.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>label</th>\n",
       "      <th>file</th>\n",
       "      <th>probability</th>\n",
       "      <th>prediction</th>\n",
       "      <th>truth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>16505</th>\n",
       "      <td>I've almost forever been against the inclusion...</td>\n",
       "      <td>0</td>\n",
       "      <td>2355_3.txt</td>\n",
       "      <td>0.907444</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18569</th>\n",
       "      <td>The beginning of this movie is excellent with ...</td>\n",
       "      <td>0</td>\n",
       "      <td>4212_4.txt</td>\n",
       "      <td>0.906162</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18682</th>\n",
       "      <td>I firmly believe that the best Oscar ceremony ...</td>\n",
       "      <td>0</td>\n",
       "      <td>4314_4.txt</td>\n",
       "      <td>0.906133</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  review  label        file  \\\n",
       "16505  I've almost forever been against the inclusion...      0  2355_3.txt   \n",
       "18569  The beginning of this movie is excellent with ...      0  4212_4.txt   \n",
       "18682  I firmly believe that the best Oscar ceremony ...      0  4314_4.txt   \n",
       "\n",
       "       probability  prediction  truth  \n",
       "16505     0.907444        True  False  \n",
       "18569     0.906162        True  False  \n",
       "18682     0.906133        True  False  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "falsenegatives = valid[(valid.truth==False)&(valid.truth!=valid.prediction)]\n",
    "print(len(falsenegatives), \"false negatives.\")\n",
    "falsenegatives.sort_values(\"probability\", ascending=False).head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "If this is based on the true-life relationship, as purported, between Ms. Curtin and Mr. Levinson, I'm thrilled I do not know them personally. This is painfully slow, and both characters take stupid pills liberally throughout the movie while the theme song gets played into the ground. Many stupid scenes with people acting stupid does not make for a comedy."
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "HTML(valid.loc[22148].review)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.095343485\n",
      "Negative!\n"
     ]
    }
   ],
   "source": [
    "unseen = pd.Series('this movie is very good')\n",
    "print(probability)\n",
    "print(\"Positive!\") if probability > 0.5 else print(\"Negative!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Transformed.\n"
     ]
    }
   ],
   "source": [
    "unseen = preprocess.transform(unseen)       # Text preprocessing\n",
    "unseen = tfidf.transform(unseen).toarray()  # Feature engineering\n",
    "unseen = unseen[:,bestIndeces]              # Feature selection\n",
    "probability = model.predict(unseen)[0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6559083\n",
      "Positive!\n"
     ]
    }
   ],
   "source": [
    "print(probability)\n",
    "print(\"Positive!\") if probability > 0.5 else print(\"Negative!\")"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}