Keras_BOW_WSD_ALL3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import itertools\n",
    "import os\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "\n",
    "import logging\n",
    "from numpy import random\n",
    "import gensim\n",
    "import nltk\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "import matplotlib.pyplot as plt\n",
    "from nltk.corpus import stopwords\n",
    "import re\n",
    "from bs4 import BeautifulSoup\n",
    "\n",
    "from sklearn.preprocessing import LabelBinarizer, LabelEncoder\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "from tensorflow import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation, Dropout\n",
    "from keras.preprocessing import text, sequence\n",
    "from keras import utils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4729\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_excel('wsd_data.xlsx')\n",
    "df = df[pd.notnull(df['CATEGORY'])]\n",
    "print(df['SENTENCE'].apply(lambda x: len(x.split(' '))).sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4729\n"
     ]
    }
   ],
   "source": [
    "def clean_text(text):\n",
    "    # Strip HTML tags\n",
    "    text = re.sub('<[^<]+?>', ' ', text)\n",
    " \n",
    "    # Strip escaped quotes\n",
    "    text = text.replace('\\\\\"', '')\n",
    " \n",
    "    # Strip quotes\n",
    "    text = text.replace('\"', '')\n",
    " \n",
    "    return text\n",
    "\n",
    "df['CLEAN_SENTENCE'] = df['SENTENCE'].apply(clean_text)\n",
    "print(df['CLEAN_SENTENCE'].apply(lambda x: len(x.split(' '))).sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "my_tags = ['manisan','manusia','kaki','tangan','periksa','tumbuhan']\n",
    "plt.figure(figsize=(10,6))\n",
    "df.CATEGORY.value_counts().plot(kind='bar');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.CLEAN_SENTENCE\n",
    "y = df.CATEGORY\n",
    "\n",
    "train_posts, test_posts, train_tags, test_tags = train_test_split(X, y, test_size=0.2, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 512)               512512    \n",
      "_________________________________________________________________\n",
      "activation_1 (Activation)    (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 6)                 3078      \n",
      "_________________________________________________________________\n",
      "activation_2 (Activation)    (None, 6)                 0         \n",
      "=================================================================\n",
      "Total params: 515,590\n",
      "Trainable params: 515,590\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "max_words = 1000\n",
    "#tokenize = text.Tokenizer(num_words=max_words, lower=True, split=' ', char_level=False, oov_token=None, document_count=0)\n",
    "tokenize = text.Tokenizer(num_words=max_words, char_level=False)\n",
    "tokenize.fit_on_texts(train_posts) # only fit on train\n",
    "\n",
    "x_train = tokenize.texts_to_matrix(train_posts)\n",
    "x_test = tokenize.texts_to_matrix(test_posts)\n",
    "\n",
    "encoder = LabelEncoder()\n",
    "encoder.fit(train_tags)\n",
    "y_train = encoder.transform(train_tags)\n",
    "y_test = encoder.transform(test_tags)\n",
    "\n",
    "num_classes = np.max(y_train) + 1\n",
    "y_train = utils.to_categorical(y_train, num_classes)\n",
    "y_test = utils.to_categorical(y_test, num_classes)\n",
    "\n",
    "batch_size = 128 #32\n",
    "epochs = 10\n",
    "\n",
    "# Build the model\n",
    "model = Sequential()\n",
    "model.add(Dense(512, input_shape=(max_words,)))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(num_classes))\n",
    "model.add(Activation('softmax'))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 169 samples, validate on 19 samples\n",
      "Epoch 1/10\n",
      "169/169 [==============================] - 3s 19ms/step - loss: 1.8270 - acc: 0.1243 - val_loss: 1.7770 - val_acc: 0.2105\n",
      "Epoch 2/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 1.6430 - acc: 0.4320 - val_loss: 1.6612 - val_acc: 0.4211\n",
      "Epoch 3/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 1.4695 - acc: 0.7337 - val_loss: 1.5538 - val_acc: 0.4211\n",
      "Epoch 4/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 1.3200 - acc: 0.8817 - val_loss: 1.4549 - val_acc: 0.6842\n",
      "Epoch 5/10\n",
      "169/169 [==============================] - 0s 370us/step - loss: 1.1916 - acc: 0.9172 - val_loss: 1.3624 - val_acc: 0.7368\n",
      "Epoch 6/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 1.0688 - acc: 0.8994 - val_loss: 1.2749 - val_acc: 0.7895\n",
      "Epoch 7/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 0.9592 - acc: 0.9349 - val_loss: 1.1922 - val_acc: 0.7895\n",
      "Epoch 8/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 0.8437 - acc: 0.9467 - val_loss: 1.1137 - val_acc: 0.7895\n",
      "Epoch 9/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 0.7695 - acc: 0.9408 - val_loss: 1.0399 - val_acc: 0.8421\n",
      "Epoch 10/10\n",
      "169/169 [==============================] - 0s 277us/step - loss: 0.6934 - acc: 0.9527 - val_loss: 0.9705 - val_acc: 0.8421\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(x_train, y_train,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_split=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot training & validation accuracy values\n",
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Plot training & validation loss values\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "48/48 [==============================] - 0s 0us/step\n",
      "Test accuracy: 0.75\n"
     ]
    }
   ],
   "source": [
    "score = model.evaluate(x_test, y_test,\n",
    "                       batch_size=batch_size, verbose=1)\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}