5- Sentiment Classifier on IMDB dataset using Conv-net.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sentiment Classifier on IMDB dataset using Conv-net\n",
    "<span style=\"color:red;\">Aims:</span><br>\n",
    "    1-Understanding NLP concept</br>\n",
    "    2-Understanding of Word2Vec Concept<br>\n",
    "    3-Understanding of Embedding,SpatialDropout1D, Conv1D, GlobalMaxPooling1D Layers<br>\n",
    "    4-Understanding of 'Area under the curve'<br>\n",
    "    5-Learn to work with ModelCheckpoint<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.datasets import imdb\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Embedding\n",
    "from keras.layers import SpatialDropout1D, Conv1D, GlobalMaxPooling1D \n",
    "from keras.callbacks import ModelCheckpoint\n",
    "import os\n",
    "from sklearn.metrics import roc_auc_score \n",
    "import matplotlib.pyplot as plt \n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_unique_words = 5000\n",
    "(x_train, y_train), (x_valid, y_valid) = imdb.load_data(num_words=n_unique_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Working with Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "max_review_length = 400\n",
    "pad_type = trunc_type = 'pre'\n",
    "x_train = pad_sequences(x_train, maxlen=max_review_length, padding=pad_type, truncating=trunc_type, value=0)\n",
    "x_valid = pad_sequences(x_valid, maxlen=max_review_length, padding=pad_type, truncating=trunc_type, value=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 400, 64)           320000    \n",
      "_________________________________________________________________\n",
      "spatial_dropout1d_1 (Spatial (None, 400, 64)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            (None, 398, 256)          49408     \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d_1 (Glob (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 256)               65792     \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 257       \n",
      "=================================================================\n",
      "Total params: 435,457\n",
      "Trainable params: 435,457\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Embedding(n_unique_words, 64, input_length=max_review_length)) \n",
    "model.add(SpatialDropout1D(0.2))\n",
    "model.add(Conv1D(256, 3, activation='relu'))\n",
    "model.add(GlobalMaxPooling1D())\n",
    "model.add(Dense(256, activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Configure model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Checkpoints"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "output_dir = 'model_output/Sent-Conv/'\n",
    "modelcheckpoint = ModelCheckpoint(filepath=output_dir+\"/weights.{epoch:02d}.hdf5\")\n",
    "if not os.path.exists(output_dir):\n",
    "    os.makedirs(output_dir)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 25000 samples, validate on 25000 samples\n",
      "Epoch 1/5\n",
      "25000/25000 [==============================] - 146s 6ms/step - loss: 0.4892 - acc: 0.7458 - val_loss: 0.2960 - val_acc: 0.8759\n",
      "Epoch 2/5\n",
      "25000/25000 [==============================] - 156s 6ms/step - loss: 0.2541 - acc: 0.8963 - val_loss: 0.2583 - val_acc: 0.8913\n",
      "Epoch 3/5\n",
      "25000/25000 [==============================] - 146s 6ms/step - loss: 0.1715 - acc: 0.9366 - val_loss: 0.2568 - val_acc: 0.8964\n",
      "Epoch 4/5\n",
      "25000/25000 [==============================] - 155s 6ms/step - loss: 0.1143 - acc: 0.9609 - val_loss: 0.2793 - val_acc: 0.8926\n",
      "Epoch 5/5\n",
      "25000/25000 [==============================] - 145s 6ms/step - loss: 0.0700 - acc: 0.9784 - val_loss: 0.3337 - val_acc: 0.8881\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x270acb2fcc0>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(x_train, y_train, batch_size=128, epochs=5, verbose=1, validation_data=(x_valid, y_valid), callbacks=[modelcheckpoint])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.load_weights(output_dir+\"weights.03.hdf5\")\n",
    "y_hat = model.predict_proba(x_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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": [
    "plt.hist(y_hat)\n",
    "_ = plt.axvline(x=0.5, color='orange')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'96.14'"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"{:0.2f}\".format(roc_auc_score(y_valid, y_hat)*100.0)"
   ]
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}