Ensemble-word-character-news-classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loading & Analysis of News Aggregator Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import keras\n",
    "import tensorflow as tf\n",
    "\n",
    "# keras: for data processing\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils.np_utils import to_categorical\n",
    "\n",
    "# keras: for Deep-learing\n",
    "from tensorflow.keras.layers import *\n",
    "from tensorflow.keras.models import *\n",
    "from tensorflow.keras.activations import *\n",
    "from tensorflow.keras.backend import *\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from tensorflow.keras.optimizers import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(41)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Data load"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('./data/uci_news_aggregator.csv', delimiter = ',', skiprows = 1,\n",
    "                   names = ['ID', 'TITLE', 'URL', 'PUBLISHER', 'CATEGORY', 'STORY',\n",
    "                            'HOSTNAME', 'TIMESTAMP'], usecols=['TITLE', 'CATEGORY'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = data.reindex(np.random.permutation(data.index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>TITLE</th>\n",
       "      <th>CATEGORY</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>51683</th>\n",
       "      <td>Cloud first, mobile first: Word, Powerpoint, a...</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>364094</th>\n",
       "      <td>This Is How a 'Go-Around' Avoids Disaster at a...</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>220239</th>\n",
       "      <td>Insider Source: Apple and Samsung Renew Settle...</td>\n",
       "      <td>t</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15418</th>\n",
       "      <td>States Asks Big Retail to Halt Tobacco Sales</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>162428</th>\n",
       "      <td>FDA Proposes New Regulations on E-Cigarettes</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>346459</th>\n",
       "      <td>Paul Mazursky, director of Bob &amp; Carol &amp; Ted &amp;...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60046</th>\n",
       "      <td>BlackBerry CEO trying to improve relationships...</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>331513</th>\n",
       "      <td>Android 5.0 Lollipop Is Sealed, No Touchwiz, S...</td>\n",
       "      <td>t</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7953</th>\n",
       "      <td>Blood test for Alzheimer's risk could be a bre...</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78829</th>\n",
       "      <td>Celebrate the return of 'Game of Thrones' with...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>288328</th>\n",
       "      <td>VIDEO: At E3, Signs That Virtual Reality's Tim...</td>\n",
       "      <td>t</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15761</th>\n",
       "      <td>Yellen and the Fed, a WSJ Briefing</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28245</th>\n",
       "      <td>Teens unite at Kick Butts Day to fight tobacco...</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>155307</th>\n",
       "      <td>Facebook Q1 earnings soar to $642 million</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>364918</th>\n",
       "      <td>Solange Lands Lucky Cover; Addresses Elevator ...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>197202</th>\n",
       "      <td>FTC: Snapchat deceived users about vanishing m...</td>\n",
       "      <td>t</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>389009</th>\n",
       "      <td>Lindsay Lohan Flaunts Beach Bod, Wears Sexy Bl...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135446</th>\n",
       "      <td>Findings offer new paths towards improved fert...</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4058</th>\n",
       "      <td>You're The Boss: Girl Scouts, LeanIn Vow To Ba...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>280371</th>\n",
       "      <td>Mila Kunis doesnt like men saying We are pregnant</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    TITLE CATEGORY\n",
       "51683   Cloud first, mobile first: Word, Powerpoint, a...        b\n",
       "364094  This Is How a 'Go-Around' Avoids Disaster at a...        b\n",
       "220239  Insider Source: Apple and Samsung Renew Settle...        t\n",
       "15418        States Asks Big Retail to Halt Tobacco Sales        m\n",
       "162428       FDA Proposes New Regulations on E-Cigarettes        m\n",
       "346459  Paul Mazursky, director of Bob & Carol & Ted &...        e\n",
       "60046   BlackBerry CEO trying to improve relationships...        b\n",
       "331513  Android 5.0 Lollipop Is Sealed, No Touchwiz, S...        t\n",
       "7953    Blood test for Alzheimer's risk could be a bre...        m\n",
       "78829   Celebrate the return of 'Game of Thrones' with...        e\n",
       "288328  VIDEO: At E3, Signs That Virtual Reality's Tim...        t\n",
       "15761                  Yellen and the Fed, a WSJ Briefing        b\n",
       "28245   Teens unite at Kick Butts Day to fight tobacco...        m\n",
       "155307          Facebook Q1 earnings soar to $642 million        b\n",
       "364918  Solange Lands Lucky Cover; Addresses Elevator ...        e\n",
       "197202  FTC: Snapchat deceived users about vanishing m...        t\n",
       "389009  Lindsay Lohan Flaunts Beach Bod, Wears Sexy Bl...        e\n",
       "135446  Findings offer new paths towards improved fert...        m\n",
       "4058    You're The Boss: Girl Scouts, LeanIn Vow To Ba...        e\n",
       "280371  Mila Kunis doesnt like men saying We are pregnant        e"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(422419, 2)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Value count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f258d099810>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data.CATEGORY.value_counts().plot(kind='bar', grid=False, figsize=(8, 4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "e    152469\n",
       "b    115967\n",
       "t    108344\n",
       "m     45639\n",
       "Name: CATEGORY, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.CATEGORY.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Construct label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "e = data[data['CATEGORY'] == 'e']#[:num_of_categories]\n",
    "b = data[data['CATEGORY'] == 'b']#[:num_of_categories]\n",
    "t = data[data['CATEGORY'] == 't']#[:num_of_categories]\n",
    "m = data[data['CATEGORY'] == 'm']#[:num_of_categories]\n",
    "\n",
    "concated = pd.concat([e,b,t,m], ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>TITLE</th>\n",
       "      <th>CATEGORY</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Paul Mazursky, director of Bob &amp; Carol &amp; Ted &amp;...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Celebrate the return of 'Game of Thrones' with...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Solange Lands Lucky Cover; Addresses Elevator ...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Lindsay Lohan Flaunts Beach Bod, Wears Sexy Bl...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>You're The Boss: Girl Scouts, LeanIn Vow To Ba...</td>\n",
       "      <td>e</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>422414</th>\n",
       "      <td>Stroke Rounds: Pre-Hypertensives at Greater Ri...</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>422415</th>\n",
       "      <td>Melanoma Skin Cancer Cases Spike</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>422416</th>\n",
       "      <td>Winnebago County begins West Nile virus season...</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>422417</th>\n",
       "      <td>World Health Organization says 147 people have...</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>422418</th>\n",
       "      <td>Young blood proven good for old brain</td>\n",
       "      <td>m</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>422419 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    TITLE CATEGORY\n",
       "0       Paul Mazursky, director of Bob & Carol & Ted &...        e\n",
       "1       Celebrate the return of 'Game of Thrones' with...        e\n",
       "2       Solange Lands Lucky Cover; Addresses Elevator ...        e\n",
       "3       Lindsay Lohan Flaunts Beach Bod, Wears Sexy Bl...        e\n",
       "4       You're The Boss: Girl Scouts, LeanIn Vow To Ba...        e\n",
       "...                                                   ...      ...\n",
       "422414  Stroke Rounds: Pre-Hypertensives at Greater Ri...        m\n",
       "422415                   Melanoma Skin Cancer Cases Spike        m\n",
       "422416  Winnebago County begins West Nile virus season...        m\n",
       "422417  World Health Organization says 147 people have...        m\n",
       "422418              Young blood proven good for old brain        m\n",
       "\n",
       "[422419 rows x 2 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "concated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "concated['LABEL'] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "concated.loc[concated['CATEGORY'] == 'e', 'LABEL'] = 0\n",
    "concated.loc[concated['CATEGORY'] == 'b', 'LABEL'] = 1\n",
    "concated.loc[concated['CATEGORY'] == 't', 'LABEL'] = 2\n",
    "concated.loc[concated['CATEGORY'] == 'm', 'LABEL'] = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "concated = concated.reindex(np.random.permutation(concated.index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "359004    2\n",
       "90474     0\n",
       "392609    3\n",
       "143914    0\n",
       "288112    2\n",
       "346156    2\n",
       "358152    2\n",
       "163396    1\n",
       "66513     0\n",
       "338916    2\n",
       "Name: LABEL, dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "concated['LABEL'][:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# one-hot encoding\n",
    "labels = to_categorical(concated['LABEL'], num_classes=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 1., 0.],\n",
       "       [1., 0., 0., 0.],\n",
       "       [0., 0., 0., 1.],\n",
       "       [1., 0., 0., 0.],\n",
       "       [0., 0., 1., 0.],\n",
       "       [0., 0., 1., 0.],\n",
       "       [0., 0., 1., 0.],\n",
       "       [0., 1., 0., 0.],\n",
       "       [1., 0., 0., 0.],\n",
       "       [0., 0., 1., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Construct Train & Test Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Embedding parameters (frequency & padding size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "most_common_alphabets = 80\n",
    "most_common_words = 11000\n",
    "char_max_len = 80\n",
    "word_max_len = 80"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Character tokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer_char = Tokenizer(num_words=most_common_alphabets, char_level=True, filters='!\"#$%&()*+,-./:;<=>?@[\\]^_`{|}~', lower=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer_char.fit_on_texts(concated['TITLE'].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "sequences_char = tokenizer_char.texts_to_sequences(concated['TITLE'].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[18,\n",
       "  2,\n",
       "  12,\n",
       "  4,\n",
       "  1,\n",
       "  11,\n",
       "  10,\n",
       "  3,\n",
       "  7,\n",
       "  14,\n",
       "  1,\n",
       "  6,\n",
       "  28,\n",
       "  14,\n",
       "  5,\n",
       "  19,\n",
       "  7,\n",
       "  10,\n",
       "  2,\n",
       "  1,\n",
       "  19,\n",
       "  7,\n",
       "  10,\n",
       "  22,\n",
       "  2,\n",
       "  12,\n",
       "  1,\n",
       "  13,\n",
       "  15,\n",
       "  9,\n",
       "  12,\n",
       "  8,\n",
       "  2,\n",
       "  12,\n",
       "  4,\n",
       "  1,\n",
       "  5,\n",
       "  18,\n",
       "  1,\n",
       "  14,\n",
       "  7,\n",
       "  10,\n",
       "  10,\n",
       "  7,\n",
       "  5,\n",
       "  9,\n",
       "  4,\n",
       "  1,\n",
       "  6,\n",
       "  13,\n",
       "  8,\n",
       "  5,\n",
       "  15,\n",
       "  17,\n",
       "  13,\n",
       "  1,\n",
       "  25,\n",
       "  11,\n",
       "  8,\n",
       "  3,\n",
       "  14,\n",
       "  14,\n",
       "  7,\n",
       "  9,\n",
       "  17,\n",
       "  25],\n",
       " [11,\n",
       "  13,\n",
       "  8,\n",
       "  7,\n",
       "  4,\n",
       "  6,\n",
       "  7,\n",
       "  9,\n",
       "  2,\n",
       "  1,\n",
       "  14,\n",
       "  11,\n",
       "  23,\n",
       "  7,\n",
       "  2,\n",
       "  1,\n",
       "  8,\n",
       "  2,\n",
       "  30,\n",
       "  5,\n",
       "  7,\n",
       "  9,\n",
       "  4,\n",
       "  1,\n",
       "  18,\n",
       "  10,\n",
       "  2,\n",
       "  2,\n",
       "  6,\n",
       "  21,\n",
       "  5,\n",
       "  5,\n",
       "  12,\n",
       "  1,\n",
       "  14,\n",
       "  3,\n",
       "  11,\n",
       "  26,\n",
       "  1,\n",
       "  9,\n",
       "  24,\n",
       "  1,\n",
       "  3,\n",
       "  14,\n",
       "  2,\n",
       "  8,\n",
       "  7,\n",
       "  11,\n",
       "  3,\n",
       "  1,\n",
       "  6,\n",
       "  5,\n",
       "  15,\n",
       "  8,\n",
       "  1,\n",
       "  4,\n",
       "  2,\n",
       "  6]]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sequences_char[:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Word tokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer_word = Tokenizer(num_words=most_common_words, char_level=False, filters='!\"#$%&()*+,-./:;<=>?@[\\]^_`{|}~', lower=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer_word.fit_on_texts(concated['TITLE'].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "sequences_word = tokenizer_word.texts_to_sequences(concated['TITLE'].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[1593, 2274, 111, 146, 2437, 4, 902, 695, 10934],\n",
       " [3454, 5039, 10733, 3611, 1232, 1821, 223, 250, 128]]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sequences_word[:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_index_char = tokenizer_char.word_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_index_word = tokenizer_word.word_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# word_index_char"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# word_index_word"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Padding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_char = pad_sequences(sequences_char, maxlen=char_max_len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_word = pad_sequences(sequences_word, maxlen=word_max_len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(422419, 80)\n",
      "(422419, 80)\n",
      "(422419, 4)\n"
     ]
    }
   ],
   "source": [
    "print(X_char.shape)\n",
    "print(X_word.shape)\n",
    "print(labels.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Character-level train & test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_char, X_test_char, y_train, y_test = train_test_split(X_char, labels, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "((337935, 80), (337935, 4), (84484, 80), (84484, 4))\n"
     ]
    }
   ],
   "source": [
    "print((X_train_char.shape, y_train.shape, X_test_char.shape, y_test.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Word-level train & test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_word, X_test_word, y_train, y_test = train_test_split(X_word, labels, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "((337935, 80), (337935, 4), (84484, 80), (84484, 4))\n"
     ]
    }
   ],
   "source": [
    "print((X_train_word.shape, y_train.shape, X_test_word.shape, y_test.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Single Classifier (character-level)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Embedding parameter (embedding vector dimension)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "char_emb_dim = 128"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_char_model():\n",
    "    inp = Input(shape=(char_max_len,))\n",
    "    emb = Embedding(most_common_alphabets, char_emb_dim, input_length=char_max_len)(inp)\n",
    "    _ = Dropout(rate=0.7)(emb)\n",
    "    _ = Conv1D(256, 3, padding='valid', activation='selu', strides=1)(_)\n",
    "    _ = MaxPooling1D()(_)\n",
    "    _ = Conv1D(128, 3, padding='valid', activation='selu', strides=1)(_)\n",
    "    _ = GlobalMaxPooling1D()(_)\n",
    "    _ = Dense(64, activation='selu')(_)\n",
    "    _ = Dropout(rate=0.7)(_)\n",
    "    out = Dense(4, activation='softmax')(_)\n",
    "    \n",
    "    return Model(inp, out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Large dropout rate: 0.7 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n",
      "WARNING:tensorflow:Large dropout rate: 0.7 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n"
     ]
    }
   ],
   "source": [
    "char_model = get_char_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 80)]              0         \n",
      "_________________________________________________________________\n",
      "embedding (Embedding)        (None, 80, 128)           10240     \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 80, 128)           0         \n",
      "_________________________________________________________________\n",
      "conv1d (Conv1D)              (None, 78, 256)           98560     \n",
      "_________________________________________________________________\n",
      "max_pooling1d (MaxPooling1D) (None, 39, 256)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            (None, 37, 128)           98432     \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d (Global (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 4)                 260       \n",
      "=================================================================\n",
      "Total params: 215,748\n",
      "Trainable params: 215,748\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "char_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Train parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 50\n",
    "batch_size = 256\n",
    "learning_rate = 0.001\n",
    "decay = 1e-06"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Compile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "char_model.compile(optimizer = Adam(lr = learning_rate, decay = decay), loss='categorical_crossentropy', metrics=['acc'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 337935 samples, validate on 84484 samples\n",
      "Epoch 1/50\n",
      "WARNING:tensorflow:Large dropout rate: 0.7 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n",
      "WARNING:tensorflow:Large dropout rate: 0.7 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n",
      "WARNING:tensorflow:Large dropout rate: 0.7 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n",
      "337935/337935 [==============================] - 32s 94us/sample - loss: 0.9618 - acc: 0.6076 - val_loss: 0.5992 - val_acc: 0.7800\n",
      "Epoch 2/50\n",
      "337935/337935 [==============================] - 31s 90us/sample - loss: 0.6949 - acc: 0.7414 - val_loss: 0.4785 - val_acc: 0.8279\n",
      "Epoch 3/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.6071 - acc: 0.7804 - val_loss: 0.4660 - val_acc: 0.8294\n",
      "Epoch 4/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.5622 - acc: 0.7991 - val_loss: 0.4020 - val_acc: 0.8555\n",
      "Epoch 5/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.5315 - acc: 0.8115 - val_loss: 0.3996 - val_acc: 0.8578\n",
      "Epoch 6/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.5100 - acc: 0.8204 - val_loss: 0.3654 - val_acc: 0.8704\n",
      "Epoch 7/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4915 - acc: 0.8278 - val_loss: 0.3531 - val_acc: 0.8743\n",
      "Epoch 8/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4761 - acc: 0.8338 - val_loss: 0.3490 - val_acc: 0.8787\n",
      "Epoch 9/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4641 - acc: 0.8381 - val_loss: 0.3359 - val_acc: 0.8811\n",
      "Epoch 10/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4521 - acc: 0.8426 - val_loss: 0.3397 - val_acc: 0.8809\n",
      "Epoch 11/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4423 - acc: 0.8453 - val_loss: 0.3341 - val_acc: 0.8837\n",
      "Epoch 12/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4305 - acc: 0.8503 - val_loss: 0.3249 - val_acc: 0.8873\n",
      "Epoch 13/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4198 - acc: 0.8544 - val_loss: 0.3191 - val_acc: 0.8898\n",
      "Epoch 14/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4122 - acc: 0.8572 - val_loss: 0.3055 - val_acc: 0.8932\n",
      "Epoch 15/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.4047 - acc: 0.8595 - val_loss: 0.2992 - val_acc: 0.8946\n",
      "Epoch 16/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3980 - acc: 0.8617 - val_loss: 0.3002 - val_acc: 0.8977\n",
      "Epoch 17/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3932 - acc: 0.8632 - val_loss: 0.3040 - val_acc: 0.8948\n",
      "Epoch 18/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3892 - acc: 0.8656 - val_loss: 0.2932 - val_acc: 0.8987\n",
      "Epoch 19/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3846 - acc: 0.8668 - val_loss: 0.2941 - val_acc: 0.8980\n",
      "Epoch 20/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3801 - acc: 0.8686 - val_loss: 0.2896 - val_acc: 0.8996\n",
      "Epoch 21/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3760 - acc: 0.8700 - val_loss: 0.2895 - val_acc: 0.9020\n",
      "Epoch 22/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3739 - acc: 0.8700 - val_loss: 0.2889 - val_acc: 0.9022\n",
      "Epoch 23/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3699 - acc: 0.8723 - val_loss: 0.2858 - val_acc: 0.9033\n",
      "Epoch 24/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3678 - acc: 0.8732 - val_loss: 0.2789 - val_acc: 0.9041\n",
      "Epoch 25/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3649 - acc: 0.8749 - val_loss: 0.2855 - val_acc: 0.9039\n",
      "Epoch 26/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3627 - acc: 0.8750 - val_loss: 0.2813 - val_acc: 0.9053\n",
      "Epoch 27/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3597 - acc: 0.8762 - val_loss: 0.2773 - val_acc: 0.9056\n",
      "Epoch 28/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3545 - acc: 0.8781 - val_loss: 0.2804 - val_acc: 0.9065\n",
      "Epoch 29/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3560 - acc: 0.8776 - val_loss: 0.2712 - val_acc: 0.9073\n",
      "Epoch 30/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3531 - acc: 0.8786 - val_loss: 0.2740 - val_acc: 0.9079\n",
      "Epoch 31/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3504 - acc: 0.8793 - val_loss: 0.2720 - val_acc: 0.9071\n",
      "Epoch 32/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3479 - acc: 0.8808 - val_loss: 0.2709 - val_acc: 0.9078\n",
      "Epoch 33/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3452 - acc: 0.8813 - val_loss: 0.2746 - val_acc: 0.9081\n",
      "Epoch 34/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3469 - acc: 0.8806 - val_loss: 0.2692 - val_acc: 0.9082\n",
      "Epoch 35/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3423 - acc: 0.8825 - val_loss: 0.2680 - val_acc: 0.9082\n",
      "Epoch 36/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3421 - acc: 0.8825 - val_loss: 0.2644 - val_acc: 0.9105\n",
      "Epoch 37/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3401 - acc: 0.8828 - val_loss: 0.2693 - val_acc: 0.9092\n",
      "Epoch 38/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3373 - acc: 0.8838 - val_loss: 0.2672 - val_acc: 0.9100\n",
      "Epoch 39/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3364 - acc: 0.8843 - val_loss: 0.2632 - val_acc: 0.9108\n",
      "Epoch 40/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3340 - acc: 0.8852 - val_loss: 0.2643 - val_acc: 0.9118\n",
      "Epoch 41/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3334 - acc: 0.8854 - val_loss: 0.2616 - val_acc: 0.9106\n",
      "Epoch 42/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3328 - acc: 0.8860 - val_loss: 0.2705 - val_acc: 0.9094\n",
      "Epoch 43/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3312 - acc: 0.8864 - val_loss: 0.2620 - val_acc: 0.9121\n",
      "Epoch 44/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3282 - acc: 0.8871 - val_loss: 0.2636 - val_acc: 0.9120\n",
      "Epoch 45/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3275 - acc: 0.8873 - val_loss: 0.2587 - val_acc: 0.9136\n",
      "Epoch 46/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3273 - acc: 0.8877 - val_loss: 0.2639 - val_acc: 0.9133\n",
      "Epoch 47/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3247 - acc: 0.8882 - val_loss: 0.2607 - val_acc: 0.9121\n",
      "Epoch 48/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3233 - acc: 0.8894 - val_loss: 0.2592 - val_acc: 0.9136\n",
      "Epoch 49/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3230 - acc: 0.8891 - val_loss: 0.2590 - val_acc: 0.9144\n",
      "Epoch 50/50\n",
      "337935/337935 [==============================] - 31s 92us/sample - loss: 0.3210 - acc: 0.8899 - val_loss: 0.2562 - val_acc: 0.9135\n"
     ]
    }
   ],
   "source": [
    "char_model_history = char_model.fit(X_train_char, y_train, epochs=epochs, batch_size=batch_size, validation_data=(X_test_char, y_test), callbacks=[EarlyStopping(monitor='val_loss',patience=7, min_delta=0.0001)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Save"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "char_model.save('char-level_model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Single Classifier (word-level)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Embedding parameter (embedding vector dimension)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_emb_dim = 128"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_word_model():\n",
    "    inp = Input(shape=(word_max_len,))\n",
    "    emb = Embedding(most_common_words, word_emb_dim, input_length=word_max_len)(inp)\n",
    "    _ = Dropout(rate=0.7)(emb)\n",
    "    _ = Conv1D(256, 3, padding='valid', activation='selu', strides=1)(_)\n",
    "    _ = MaxPooling1D()(_)\n",
    "    _ = Conv1D(128, 3, padding='valid', activation='selu', strides=1)(_)\n",
    "    _ = GlobalMaxPooling1D()(_)\n",
    "    _ = Dense(64, activation='selu')(_)\n",
    "    _ = Dropout(rate=0.7)(_)\n",
    "    out = Dense(4, activation='softmax')(_)\n",
    "    \n",
    "    return Model(inp, out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_model = get_word_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (InputLayer)         [(None, 80)]              0         \n",
      "_________________________________________________________________\n",
      "embedding_1 (Embedding)      (None, 80, 128)           1408000   \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 80, 128)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_2 (Conv1D)            (None, 78, 256)           98560     \n",
      "_________________________________________________________________\n",
      "max_pooling1d_1 (MaxPooling1 (None, 39, 256)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_3 (Conv1D)            (None, 37, 128)           98432     \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d_1 (Glob (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 4)                 260       \n",
      "=================================================================\n",
      "Total params: 1,613,508\n",
      "Trainable params: 1,613,508\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "word_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Train parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 50\n",
    "batch_size = 256\n",
    "learning_rate = 0.001\n",
    "decay = 1e-06"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Compile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_model.compile(optimizer = Adam(lr = learning_rate, decay = decay), loss='categorical_crossentropy', metrics=['acc'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 337935 samples, validate on 84484 samples\n",
      "Epoch 1/50\n",
      "337935/337935 [==============================] - 38s 114us/sample - loss: 0.3701 - acc: 0.8650 - val_loss: 0.1950 - val_acc: 0.9337\n",
      "Epoch 2/50\n",
      "337935/337935 [==============================] - 37s 110us/sample - loss: 0.2148 - acc: 0.9293 - val_loss: 0.1820 - val_acc: 0.9378\n",
      "Epoch 3/50\n",
      "337935/337935 [==============================] - 37s 110us/sample - loss: 0.1956 - acc: 0.9353 - val_loss: 0.1792 - val_acc: 0.9389\n",
      "Epoch 4/50\n",
      "337935/337935 [==============================] - 37s 110us/sample - loss: 0.1842 - acc: 0.9388 - val_loss: 0.1749 - val_acc: 0.9401\n",
      "Epoch 5/50\n",
      "337935/337935 [==============================] - 37s 110us/sample - loss: 0.1767 - acc: 0.9410 - val_loss: 0.1766 - val_acc: 0.9406\n",
      "Epoch 6/50\n",
      "337935/337935 [==============================] - 37s 110us/sample - loss: 0.1696 - acc: 0.9430 - val_loss: 0.1745 - val_acc: 0.9416\n",
      "Epoch 7/50\n",
      "337935/337935 [==============================] - 37s 110us/sample - loss: 0.1645 - acc: 0.9451 - val_loss: 0.1747 - val_acc: 0.9419\n",
      "Epoch 8/50\n",
      "337935/337935 [==============================] - 37s 110us/sample - loss: 0.1601 - acc: 0.9466 - val_loss: 0.1748 - val_acc: 0.9410\n",
      "Epoch 9/50\n",
      "337935/337935 [==============================] - 37s 111us/sample - loss: 0.1548 - acc: 0.9478 - val_loss: 0.1763 - val_acc: 0.9422\n",
      "Epoch 10/50\n",
      "337935/337935 [==============================] - 37s 111us/sample - loss: 0.1523 - acc: 0.9487 - val_loss: 0.1786 - val_acc: 0.9431\n",
      "Epoch 11/50\n",
      "337935/337935 [==============================] - 37s 111us/sample - loss: 0.1478 - acc: 0.9502 - val_loss: 0.1808 - val_acc: 0.9421\n",
      "Epoch 12/50\n",
      "337935/337935 [==============================] - 38s 111us/sample - loss: 0.1441 - acc: 0.9508 - val_loss: 0.1810 - val_acc: 0.9429\n",
      "Epoch 13/50\n",
      "337935/337935 [==============================] - 38s 111us/sample - loss: 0.1411 - acc: 0.9524 - val_loss: 0.1828 - val_acc: 0.9427\n"
     ]
    }
   ],
   "source": [
    "word_model_history = word_model.fit(X_train_word, y_train, epochs=epochs, batch_size=batch_size, validation_data=(X_test_word, y_test), callbacks=[EarlyStopping(monitor='val_loss',patience=7, min_delta=0.0001)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Save"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_model.save('word-level_model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Ensemble Model (character-level + word-level)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Load model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "load_char_model = load_model('char-level_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 80)]              0         \n",
      "_________________________________________________________________\n",
      "embedding (Embedding)        (None, 80, 128)           10240     \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 80, 128)           0         \n",
      "_________________________________________________________________\n",
      "conv1d (Conv1D)              (None, 78, 256)           98560     \n",
      "_________________________________________________________________\n",
      "max_pooling1d (MaxPooling1D) (None, 39, 256)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            (None, 37, 128)           98432     \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d (Global (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 4)                 260       \n",
      "=================================================================\n",
      "Total params: 215,748\n",
      "Trainable params: 215,748\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "load_char_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "load_word_model = load_model('word_model_conv1d_final.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_3\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_4 (InputLayer)         [(None, 80)]              0         \n",
      "_________________________________________________________________\n",
      "embedding_3 (Embedding)      (None, 80, 128)           1408000   \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 80, 128)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_5 (Conv1D)            (None, 78, 256)           98560     \n",
      "_________________________________________________________________\n",
      "max_pooling1d_3 (MaxPooling1 (None, 39, 256)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_6 (Conv1D)            (None, 37, 128)           98432     \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d_3 (Glob (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 4)                 260       \n",
      "=================================================================\n",
      "Total params: 1,613,508\n",
      "Trainable params: 1,613,508\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "load_word_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Ensemble model (ensemble of each single model using character-level and word-level feature space)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_ensemble_model():\n",
    "    \n",
    "    concat_layer = concatenate([load_char_model.layers[-2].output, load_word_model.layers[-2].output], axis=-1)\n",
    "    \n",
    "    _ = Dense(64, activation='selu', name='gate_linear')(concat_layer)\n",
    "    _ = Dropout(0.8, name='gate_dropout')(_)\n",
    "\n",
    "    out = Dense(4, activation='softmax', name='SoftMax')(_)\n",
    "    \n",
    "    return Model([load_char_model.input,load_word_model.input], out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_char_model = get_ensemble_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_2\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            [(None, 80)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_4 (InputLayer)            [(None, 80)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "embedding (Embedding)           (None, 80, 128)      10240       input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "embedding_3 (Embedding)         (None, 80, 128)      1408000     input_4[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dropout (Dropout)               (None, 80, 128)      0           embedding[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dropout_5 (Dropout)             (None, 80, 128)      0           embedding_3[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "conv1d (Conv1D)                 (None, 78, 256)      98560       dropout[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "conv1d_5 (Conv1D)               (None, 78, 256)      98560       dropout_5[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling1d (MaxPooling1D)    (None, 39, 256)      0           conv1d[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling1d_3 (MaxPooling1D)  (None, 39, 256)      0           conv1d_5[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv1d_1 (Conv1D)               (None, 37, 128)      98432       max_pooling1d[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv1d_6 (Conv1D)               (None, 37, 128)      98432       max_pooling1d_3[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "global_max_pooling1d (GlobalMax (None, 128)          0           conv1d_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "global_max_pooling1d_3 (GlobalM (None, 128)          0           conv1d_6[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense (Dense)                   (None, 64)           8256        global_max_pooling1d[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "dense_5 (Dense)                 (None, 64)           8256        global_max_pooling1d_3[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)             (None, 64)           0           dense[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "dropout_6 (Dropout)             (None, 64)           0           dense_5[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_concat (TensorFlowO [(None, 128)]        0           dropout_1[0][0]                  \n",
      "                                                                 dropout_6[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "gate_linear (Dense)             (None, 64)           8256        tf_op_layer_concat[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "gate_dropout (Dropout)          (None, 64)           0           gate_linear[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "SoftMax (Dense)                 (None, 4)            260         gate_dropout[0][0]               \n",
      "==================================================================================================\n",
      "Total params: 1,837,252\n",
      "Trainable params: 1,837,252\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "word_char_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Train parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 20\n",
    "batch_size = 256\n",
    "learning_rate = 0.001\n",
    "decay = 1e-06"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Compile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_char_model.compile(optimizer = Adam(lr = learning_rate, decay = decay), loss='categorical_crossentropy', metrics=['acc'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 337935 samples, validate on 84484 samples\n",
      "Epoch 1/20\n",
      "337935/337935 [==============================] - 59s 175us/sample - loss: 0.8553 - acc: 0.8128 - val_loss: 0.2567 - val_acc: 0.9175\n",
      "Epoch 2/20\n",
      "337935/337935 [==============================] - 58s 171us/sample - loss: 0.3770 - acc: 0.8803 - val_loss: 0.2545 - val_acc: 0.9195\n",
      "Epoch 3/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.3469 - acc: 0.8887 - val_loss: 0.2401 - val_acc: 0.9244\n",
      "Epoch 4/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.3057 - acc: 0.9030 - val_loss: 0.2164 - val_acc: 0.9339\n",
      "Epoch 5/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.2686 - acc: 0.9154 - val_loss: 0.2006 - val_acc: 0.9393\n",
      "Epoch 6/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.2450 - acc: 0.9235 - val_loss: 0.1950 - val_acc: 0.9392\n",
      "Epoch 7/20\n",
      "337935/337935 [==============================] - 58s 171us/sample - loss: 0.2286 - acc: 0.9280 - val_loss: 0.1865 - val_acc: 0.9418\n",
      "Epoch 8/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.2170 - acc: 0.9321 - val_loss: 0.1849 - val_acc: 0.9438\n",
      "Epoch 9/20\n",
      "337935/337935 [==============================] - 58s 171us/sample - loss: 0.2077 - acc: 0.9350 - val_loss: 0.1792 - val_acc: 0.9444\n",
      "Epoch 10/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.1997 - acc: 0.9374 - val_loss: 0.1828 - val_acc: 0.9453\n",
      "Epoch 11/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.1951 - acc: 0.9388 - val_loss: 0.1796 - val_acc: 0.9447\n",
      "Epoch 12/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.1907 - acc: 0.9402 - val_loss: 0.1920 - val_acc: 0.9451\n",
      "Epoch 13/20\n",
      "337935/337935 [==============================] - 58s 171us/sample - loss: 0.1871 - acc: 0.9412 - val_loss: 0.1822 - val_acc: 0.9453\n",
      "Epoch 14/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.1812 - acc: 0.9428 - val_loss: 0.1844 - val_acc: 0.9459\n",
      "Epoch 15/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.1794 - acc: 0.9431 - val_loss: 0.1882 - val_acc: 0.9464\n",
      "Epoch 16/20\n",
      "337935/337935 [==============================] - 58s 172us/sample - loss: 0.1745 - acc: 0.9447 - val_loss: 0.1886 - val_acc: 0.9464\n"
     ]
    }
   ],
   "source": [
    "word_char_model_history = word_char_model.fit([X_train_char, X_train_word], y_train, epochs=epochs, batch_size=batch_size, validation_data=([X_test_char, X_test_word], y_test), callbacks=[EarlyStopping(monitor='val_loss',patience=7, min_delta=0.0001)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Save"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_char_model.save('word_char_ensemble_model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ROC curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import label_binarize\n",
    "from sklearn.multiclass import OneVsRestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_true = np.array(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_num = 4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Character-level"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = np.array(load_char_model.predict(X_test_char))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 1., 0.],\n",
       "       [0., 1., 0., 0.],\n",
       "       [0., 0., 1., 0.],\n",
       "       ...,\n",
       "       [1., 0., 0., 0.],\n",
       "       [0., 1., 0., 0.],\n",
       "       [0., 1., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_true"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[2.1298766e-01, 4.5621369e-02, 7.1290720e-01, 2.8483834e-02],\n",
       "       [1.0089307e-01, 6.1738902e-01, 1.4550908e-01, 1.3620880e-01],\n",
       "       [1.1689332e-03, 1.6017379e-03, 9.9703985e-01, 1.8950972e-04],\n",
       "       ...,\n",
       "       [9.9995470e-01, 2.7149808e-05, 1.5951389e-05, 2.1666726e-06],\n",
       "       [1.1454013e-01, 4.4927493e-01, 3.6502454e-01, 7.1160384e-02],\n",
       "       [6.3688740e-02, 3.6916718e-01, 2.5262436e-01, 3.1451973e-01]],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(84484, 4)\n",
      "(84484, 4)\n"
     ]
    }
   ],
   "source": [
    "print(y_pred.shape)\n",
    "print(y_true.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr = dict()\n",
    "tpr = dict()\n",
    "roc_auc = dict()\n",
    "\n",
    "for i in range(class_num):\n",
    "#     print(y_true[:, i])\n",
    "#     print(y_pred[:, i])\n",
    "    \n",
    "    fpr[i], tpr[i], _ = roc_curve(y_true[:, i], y_pred[:, i])\n",
    "    \n",
    "#     print(fpr)\n",
    "#     print(tpr)\n",
    "\n",
    "    roc_auc[i] = auc(fpr[i], tpr[i])\n",
    "    eer_threshold = fpr[i][np.argmin(np.abs(fpr[i] + tpr[i] - 1))]\n",
    "    \n",
    "    plt.figure()\n",
    "    lw = 2\n",
    "    plt.plot(fpr[i], tpr[i], color='darkorange',\n",
    "             lw=lw, label='area = %0.06f, EER: %0.06f' % (roc_auc[i], eer_threshold))\n",
    "    plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.title('class ' + str(i + 1))\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "char_all_fpr = np.unique(np.concatenate([fpr[i] for i in range(class_num)]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.00000000e+00, 1.32708718e-05, 1.59360010e-05, ...,\n",
       "       9.99057768e-01, 9.99084310e-01, 1.00000000e+00])"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "char_all_fpr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/taeyong/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:5: DeprecationWarning: scipy.interp is deprecated and will be removed in SciPy 2.0.0, use numpy.interp instead\n",
      "  \"\"\"\n"
     ]
    }
   ],
   "source": [
    "from scipy import interp\n",
    "\n",
    "char_mean_tpr = np.zeros_like(char_all_fpr)\n",
    "for i in range(class_num):\n",
    "    char_mean_tpr += interp(char_all_fpr, fpr[i], tpr[i])\n",
    "    \n",
    "char_mean_tpr /= class_num"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.08506918, 0.09288403, 0.09786129, ..., 1.        , 1.        ,\n",
       "       1.        ])"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "char_mean_tpr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from itertools import cycle\n",
    "\n",
    "# Plot all ROC curves\n",
    "plt.figure()\n",
    "plt.plot(char_all_fpr, char_mean_tpr,\n",
    "         label='micro-average ROC curve (area = {0:0.4f})'\n",
    "               ''.format(auc(char_all_fpr, char_mean_tpr)),\n",
    "         color='deeppink', linestyle=':', linewidth=4)\n",
    "\n",
    "colors = cycle(['aqua', 'darkorange', 'cornflowerblue', 'red', 'green', 'yellow', 'gray', 'blue', 'orange', 'black'])\n",
    "for i, color in zip(range(class_num), colors):\n",
    "    category = ''\n",
    "    if i == 0: category = 'e'\n",
    "    elif i == 1: category = 'b'\n",
    "    elif i == 2: category = 't'\n",
    "    elif i == 3: category = 'm'\n",
    "    plt.plot(fpr[i], tpr[i], color=color, lw=lw,\n",
    "             label='ROC curve of category {0} (area = {1:0.4f})'\n",
    "             ''.format(category, roc_auc[i]))\n",
    "\n",
    "plt.plot([0, 1], [0, 1], 'k--', lw=lw)\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve of character-level single classifier')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- word-level"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = np.array(load_word_model.predict(X_test_word))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 1., 0.],\n",
       "       [0., 1., 0., 0.],\n",
       "       [0., 0., 1., 0.],\n",
       "       ...,\n",
       "       [1., 0., 0., 0.],\n",
       "       [0., 1., 0., 0.],\n",
       "       [0., 1., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_true"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.9278275e-02, 2.2659609e-02, 9.3044353e-01, 2.7618570e-02],\n",
       "       [6.2629728e-09, 9.9996185e-01, 3.8035720e-05, 6.4116328e-08],\n",
       "       [9.1799702e-12, 1.4894830e-06, 9.9999857e-01, 1.7707780e-08],\n",
       "       ...,\n",
       "       [9.2224371e-01, 1.8606162e-02, 5.2954923e-02, 6.1952271e-03],\n",
       "       [9.8925397e-05, 9.9494803e-01, 4.6706423e-03, 2.8244426e-04],\n",
       "       [3.6781609e-09, 9.9997115e-01, 2.8801278e-05, 7.4034979e-08]],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(84484, 4)\n",
      "(84484, 4)\n"
     ]
    }
   ],
   "source": [
    "print(y_pred.shape)\n",
    "print(y_true.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr = dict()\n",
    "tpr = dict()\n",
    "roc_auc = dict()\n",
    "\n",
    "for i in range(class_num):\n",
    "#     print(y_true[:, i])\n",
    "#     print(y_pred[:, i])\n",
    "    \n",
    "    fpr[i], tpr[i], _ = roc_curve(y_true[:, i], y_pred[:, i])\n",
    "    \n",
    "#     print(fpr)\n",
    "#     print(tpr)\n",
    "\n",
    "    roc_auc[i] = auc(fpr[i], tpr[i])\n",
    "    eer_threshold = fpr[i][np.argmin(np.abs(fpr[i] + tpr[i] - 1))]\n",
    "    \n",
    "    plt.figure()\n",
    "    lw = 2\n",
    "    plt.plot(fpr[i], tpr[i], color='darkorange',\n",
    "             lw=lw, label='area = %0.06f, EER: %0.06f' % (roc_auc[i], eer_threshold))\n",
    "    plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.title('class ' + str(i + 1))\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_all_fpr = np.unique(np.concatenate([fpr[i] for i in range(class_num)]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.00000000e+00, 1.32708718e-05, 1.59360010e-05, ...,\n",
       "       9.99389149e-01, 9.99426170e-01, 1.00000000e+00])"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_all_fpr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/taeyong/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:5: DeprecationWarning: scipy.interp is deprecated and will be removed in SciPy 2.0.0, use numpy.interp instead\n",
      "  \"\"\"\n"
     ]
    }
   ],
   "source": [
    "from scipy import interp\n",
    "\n",
    "word_mean_tpr = np.zeros_like(word_all_fpr)\n",
    "for i in range(class_num):\n",
    "    word_mean_tpr += interp(word_all_fpr, fpr[i], tpr[i])\n",
    "    \n",
    "word_mean_tpr /= class_num"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.06633901, 0.14786572, 0.1545477 , ..., 1.        , 1.        ,\n",
       "       1.        ])"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_mean_tpr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from itertools import cycle\n",
    "\n",
    "# Plot all ROC curves\n",
    "plt.figure()\n",
    "plt.plot(word_all_fpr, word_mean_tpr,\n",
    "         label='micro-average ROC curve (area = {0:0.4f})'\n",
    "               ''.format(auc(word_all_fpr, word_mean_tpr)),\n",
    "         color='deeppink', linestyle=':', linewidth=4)\n",
    "\n",
    "colors = cycle(['aqua', 'darkorange', 'cornflowerblue', 'red', 'green', 'yellow', 'gray', 'blue', 'orange', 'black'])\n",
    "for i, color in zip(range(class_num), colors):\n",
    "    category = ''\n",
    "    if i == 0: category = 'e'\n",
    "    elif i == 1: category = 'b'\n",
    "    elif i == 2: category = 't'\n",
    "    elif i == 3: category = 'm'\n",
    "    plt.plot(fpr[i], tpr[i], color=color, lw=lw,\n",
    "             label='ROC curve of category {0} (area = {1:0.4f})'\n",
    "             ''.format(category, roc_auc[i]))\n",
    "\n",
    "plt.plot([0, 1], [0, 1], 'k--', lw=lw)\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve of character-level single classifier')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Ensemble"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_char_model = load_model('word_char_ensemble_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = np.array(word_char_model.predict([X_test_char, X_test_word]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr = dict()\n",
    "tpr = dict()\n",
    "roc_auc = dict()\n",
    "\n",
    "for i in range(class_num):\n",
    "#     print(y_true[:, i])\n",
    "#     print(y_pred[:, i])\n",
    "    \n",
    "    fpr[i], tpr[i], _ = roc_curve(y_true[:, i], y_pred[:, i])\n",
    "    \n",
    "#     print(fpr)\n",
    "#     print(tpr)\n",
    "\n",
    "    roc_auc[i] = auc(fpr[i], tpr[i])\n",
    "    eer_threshold = fpr[i][np.argmin(np.abs(fpr[i] + tpr[i] - 1))]\n",
    "    \n",
    "    plt.figure()\n",
    "    lw = 2\n",
    "    plt.plot(fpr[i], tpr[i], color='darkorange',\n",
    "             lw=lw, label='area = %0.06f, EER: %0.06f' % (roc_auc[i], eer_threshold))\n",
    "    plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.title('class ' + str(i + 1))\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_char_all_fpr = np.unique(np.concatenate([fpr[i] for i in range(class_num)]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.00000000e+00, 1.63065634e-05, 2.65417435e-05, ...,\n",
       "       9.97649149e-01, 9.97686171e-01, 1.00000000e+00])"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_char_all_fpr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/taeyong/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:5: DeprecationWarning: scipy.interp is deprecated and will be removed in SciPy 2.0.0, use numpy.interp instead\n",
      "  \"\"\"\n"
     ]
    }
   ],
   "source": [
    "from scipy import interp\n",
    "\n",
    "word_char_mean_tpr = np.zeros_like(word_char_all_fpr)\n",
    "for i in range(class_num):\n",
    "    word_char_mean_tpr += interp(word_char_all_fpr, fpr[i], tpr[i])\n",
    "    \n",
    "word_char_mean_tpr /= class_num"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.04080487, 0.11919655, 0.16384081, ..., 1.        , 1.        ,\n",
       "       1.        ])"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_char_mean_tpr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from itertools import cycle\n",
    "\n",
    "# Plot all ROC curves\n",
    "plt.figure()\n",
    "plt.plot(word_char_all_fpr, word_char_mean_tpr,\n",
    "         label='micro-average ROC curve (area = {0:0.4f})'\n",
    "               ''.format(auc(word_char_all_fpr, word_char_mean_tpr)),\n",
    "         color='deeppink', linestyle=':', linewidth=4)\n",
    "\n",
    "colors = cycle(['aqua', 'darkorange', 'cornflowerblue', 'red', 'green', 'yellow', 'gray', 'blue', 'orange', 'black'])\n",
    "for i, color in zip(range(class_num), colors):\n",
    "    category = ''\n",
    "    if i == 0: category = 'e'\n",
    "    elif i == 1: category = 'b'\n",
    "    elif i == 2: category = 't'\n",
    "    elif i == 3: category = 'm'\n",
    "    plt.plot(fpr[i], tpr[i], color=color, lw=lw,\n",
    "             label='ROC curve of category {0} (area = {1:0.4f})'\n",
    "             ''.format(category, roc_auc[i]))\n",
    "\n",
    "plt.plot([0, 1], [0, 1], 'k--', lw=lw)\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve of ensemble model')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Comparison with ROC curve of each model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from itertools import cycle\n",
    "\n",
    "# Plot all ROC curves\n",
    "plt.figure()\n",
    "plt.plot(char_all_fpr, char_mean_tpr,\n",
    "         label='low-level model average ROC curve (area = {0:0.4f})'\n",
    "               ''.format(auc(char_all_fpr, char_mean_tpr)),\n",
    "         color='green', linestyle=':', linewidth=4)\n",
    "\n",
    "plt.plot(word_all_fpr, word_mean_tpr,\n",
    "         label='high-level model average ROC curve (area = {0:0.4f})'\n",
    "               ''.format(auc(word_all_fpr, word_mean_tpr)),\n",
    "         color='aqua', linestyle=':', linewidth=4)\n",
    "\n",
    "plt.plot(word_char_all_fpr, word_char_mean_tpr,\n",
    "         label='emsemble model average ROC curve (area = {0:0.4f})'\n",
    "               ''.format(auc(word_char_all_fpr, word_char_mean_tpr)),\n",
    "         color='black', linestyle=':', linewidth=4)\n",
    "\n",
    "plt.plot([0, 1], [0, 1], 'k--', lw=lw)\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Confusion Matrix "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_char_model = load_model('word_char_ensemble_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, auc\n",
    "from scipy import interp\n",
    "from itertools import cycle\n",
    "import itertools\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.utils.multiclass import unique_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "def cm_analysis(y_true, y_pred, labels, ymap=None, figsize=(10,10)):\n",
    "    \"\"\"\n",
    "    Generate matrix plot of confusion matrix with pretty annotations.\n",
    "    The plot image is saved to disk.\n",
    "    args: \n",
    "      y_true:    true label of the data, with shape (nsamples,)\n",
    "      y_pred:    prediction of the data, with shape (nsamples,)\n",
    "      filename:  filename of figure file to save\n",
    "      labels:    string array, name the order of class labels in the confusion matrix.\n",
    "                 use `clf.classes_` if using scikit-learn models.\n",
    "                 with shape (nclass,).\n",
    "      ymap:      dict: any -> string, length == nclass.\n",
    "                 if not None, map the labels & ys to more understandable strings.\n",
    "                 Caution: original y_true, y_pred and labels must align.\n",
    "      figsize:   the size of the figure plotted.\n",
    "    \"\"\"\n",
    "    if ymap is not None:\n",
    "        y_pred = [ymap[yi] for yi in y_pred]\n",
    "        y_true = [ymap[yi] for yi in y_true]\n",
    "        labels = [ymap[yi] for yi in labels]\n",
    "    cm = confusion_matrix(y_true, y_pred)\n",
    "    cm_sum = np.sum(cm, axis=1, keepdims=True)\n",
    "    cm_perc = cm / cm_sum.astype(float) * 100\n",
    "    annot = np.empty_like(cm).astype(str)\n",
    "    nrows, ncols = cm.shape\n",
    "    for i in range(nrows):\n",
    "        for j in range(ncols):\n",
    "            c = cm[i, j]\n",
    "            p = cm_perc[i, j]\n",
    "            if i == j:\n",
    "                s = cm_sum[i]\n",
    "                annot[i, j] = '%.1f%%\\n%d/%d' % (p, c, s)\n",
    "            elif c == 0:\n",
    "                annot[i, j] = ''\n",
    "            else:\n",
    "                annot[i, j] = '%.1f%%\\n%d' % (p, c)\n",
    "    cm = pd.DataFrame(cm, index=labels, columns=labels)\n",
    "    cm.index.name = 'Actual'\n",
    "    cm.columns.name = 'Predicted'\n",
    "    fig, ax = plt.subplots(figsize=figsize)\n",
    "    sns.set(font_scale=2)\n",
    "    sns.heatmap(cm, annot=annot, annot_kws={\"size\": 20}, fmt='', ax=ax, cmap=plt.cm.Blues, xticklabels=labels, yticklabels=labels)\n",
    "    #plt.savefig(filename)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_names = ['e', 'b', 't', 'm']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = word_char_model.predict([X_test_char, X_test_word])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(84484, 4)"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 1, 2, ..., 0, 1, 1])"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(y_test, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 1, 2, ..., 0, 1, 1])"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(predictions, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_analysis(np.argmax(y_test, axis=1), np.argmax(predictions, axis=1), class_names, ymap=None, figsize=(15,15))"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}