Accident_RNN_Lstm.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "curses is not supported on this machine (please install/reinstall curses for an optimal experience)\n",
      "904\n",
      "302\n",
      "904\n",
      "302\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\anaconda\\lib\\site-packages\\sklearn\\cross_validation.py:41: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 150, 32)           220608    \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                (None, 128)               82432     \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 128)               16512     \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 129       \n",
      "=================================================================\n",
      "Total params: 319,681\n",
      "Trainable params: 319,681\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from tflearn.data_utils import to_categorical, pad_sequences\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense,Embedding,LSTM,Dropout,Conv1D\n",
    "from keras.wrappers.scikit_learn import KerasClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "import pandas as pd\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from __future__ import print_function\n",
    "X=pd.read_csv(\"C:/Users/DEEPANSH/Desktop/Tweets1.csv\")\n",
    "x=X.Tweets\n",
    "y=X.Labels\n",
    "raw=x.tolist()\n",
    "i=1;\n",
    "final=[]\n",
    "dic={'address':1}\n",
    "for dil in raw:\n",
    "    l=[]\n",
    "    w=dil.split()\n",
    "    for wo in w:\n",
    "        if wo in dic.keys():\n",
    "            l.append(dic.get(wo))\n",
    "        else:\n",
    "            i=i+1\n",
    "            dic[wo]=i\n",
    "            l.append(dic.get(wo))\n",
    "    final.append(l)\n",
    "final\n",
    "Y=y\n",
    "from sklearn.cross_validation import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(final, Y ,random_state=1)\n",
    "print(len(X_train))\n",
    "print(len(X_test))\n",
    "print(len(y_train))\n",
    "print(len(y_test))\n",
    "# load the dataset but only keep the top n words, zero the rest\n",
    "from keras.preprocessing import sequence\n",
    "top_words = i+1\n",
    "# pad dataset to a maximum review length in words\n",
    "max_words = 150\n",
    "X_train = sequence.pad_sequences(X_train, maxlen=max_words)\n",
    "X_test = sequence.pad_sequences(X_test, maxlen=max_words)\n",
    "model = Sequential()\n",
    "model.add(Embedding(top_words, 32, input_length=max_words))\n",
    "model.add(LSTM(128,dropout=0.8))\n",
    "model.add(Dense(128, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 904 samples, validate on 302 samples\n",
      "Epoch 1/50\n",
      " - 3s - loss: 0.6924 - acc: 0.5321 - val_loss: 0.6902 - val_acc: 0.5530\n",
      "Epoch 2/50\n",
      " - 2s - loss: 0.6876 - acc: 0.5332 - val_loss: 0.6823 - val_acc: 0.5066\n",
      "Epoch 3/50\n",
      " - 2s - loss: 0.6703 - acc: 0.5796 - val_loss: 0.6504 - val_acc: 0.6921\n",
      "Epoch 4/50\n",
      " - 2s - loss: 0.6241 - acc: 0.7423 - val_loss: 0.5880 - val_acc: 0.7152\n",
      "Epoch 5/50\n",
      " - 2s - loss: 0.5578 - acc: 0.7157 - val_loss: 0.5211 - val_acc: 0.7450\n",
      "Epoch 6/50\n",
      " - 2s - loss: 0.4915 - acc: 0.7898 - val_loss: 0.5042 - val_acc: 0.7483\n",
      "Epoch 7/50\n",
      " - 2s - loss: 0.4155 - acc: 0.8131 - val_loss: 0.4891 - val_acc: 0.7351\n",
      "Epoch 8/50\n",
      " - 2s - loss: 0.3242 - acc: 0.8728 - val_loss: 0.4322 - val_acc: 0.8013\n",
      "Epoch 9/50\n",
      " - 2s - loss: 0.2849 - acc: 0.8894 - val_loss: 0.4206 - val_acc: 0.8113\n",
      "Epoch 10/50\n",
      " - 2s - loss: 0.2731 - acc: 0.8816 - val_loss: 0.4591 - val_acc: 0.7781\n",
      "Epoch 11/50\n",
      " - 2s - loss: 0.2090 - acc: 0.9270 - val_loss: 0.4848 - val_acc: 0.7848\n",
      "Epoch 12/50\n",
      " - 2s - loss: 0.1870 - acc: 0.9259 - val_loss: 0.4517 - val_acc: 0.7781\n",
      "Epoch 13/50\n",
      " - 2s - loss: 0.1550 - acc: 0.9458 - val_loss: 0.5192 - val_acc: 0.8013\n",
      "Epoch 14/50\n",
      " - 2s - loss: 0.1080 - acc: 0.9624 - val_loss: 0.5548 - val_acc: 0.8146\n",
      "Epoch 15/50\n",
      " - 2s - loss: 0.1455 - acc: 0.9524 - val_loss: 0.5002 - val_acc: 0.8212\n",
      "Epoch 16/50\n",
      " - 2s - loss: 0.0989 - acc: 0.9746 - val_loss: 0.4886 - val_acc: 0.8212\n",
      "Epoch 17/50\n",
      " - 2s - loss: 0.0893 - acc: 0.9757 - val_loss: 0.4981 - val_acc: 0.8245\n",
      "Epoch 18/50\n",
      " - 2s - loss: 0.0818 - acc: 0.9735 - val_loss: 0.5407 - val_acc: 0.7914\n",
      "Epoch 19/50\n",
      " - 2s - loss: 0.0789 - acc: 0.9790 - val_loss: 0.5522 - val_acc: 0.8179\n",
      "Epoch 20/50\n",
      " - 2s - loss: 0.0552 - acc: 0.9812 - val_loss: 0.6453 - val_acc: 0.8245\n",
      "Epoch 21/50\n",
      " - 2s - loss: 0.0871 - acc: 0.9746 - val_loss: 0.6896 - val_acc: 0.7914\n",
      "Epoch 22/50\n",
      " - 2s - loss: 0.0839 - acc: 0.9723 - val_loss: 0.7950 - val_acc: 0.7682\n",
      "Epoch 23/50\n",
      " - 2s - loss: 0.0772 - acc: 0.9757 - val_loss: 0.5904 - val_acc: 0.8212\n",
      "Epoch 24/50\n",
      " - 2s - loss: 0.0594 - acc: 0.9823 - val_loss: 0.5574 - val_acc: 0.7980\n",
      "Epoch 25/50\n",
      " - 2s - loss: 0.0480 - acc: 0.9878 - val_loss: 0.5704 - val_acc: 0.8245\n",
      "Epoch 26/50\n",
      " - 2s - loss: 0.0583 - acc: 0.9834 - val_loss: 0.6509 - val_acc: 0.8311\n",
      "Epoch 27/50\n",
      " - 2s - loss: 0.0650 - acc: 0.9812 - val_loss: 0.6334 - val_acc: 0.8179\n",
      "Epoch 28/50\n",
      " - 2s - loss: 0.0425 - acc: 0.9889 - val_loss: 0.5956 - val_acc: 0.8212\n",
      "Epoch 29/50\n",
      " - 2s - loss: 0.0492 - acc: 0.9889 - val_loss: 0.6669 - val_acc: 0.7881\n",
      "Epoch 30/50\n",
      " - 2s - loss: 0.0500 - acc: 0.9856 - val_loss: 0.5899 - val_acc: 0.8212\n",
      "Epoch 31/50\n",
      " - 2s - loss: 0.0382 - acc: 0.9900 - val_loss: 0.6824 - val_acc: 0.8146\n",
      "Epoch 32/50\n",
      " - 2s - loss: 0.0397 - acc: 0.9889 - val_loss: 0.7043 - val_acc: 0.8344\n",
      "Epoch 33/50\n",
      " - 2s - loss: 0.0418 - acc: 0.9889 - val_loss: 0.7122 - val_acc: 0.7914\n",
      "Epoch 34/50\n",
      " - 2s - loss: 0.0324 - acc: 0.9923 - val_loss: 0.7800 - val_acc: 0.7914\n",
      "Epoch 35/50\n",
      " - 2s - loss: 0.0421 - acc: 0.9900 - val_loss: 0.7435 - val_acc: 0.7748\n",
      "Epoch 36/50\n",
      " - 2s - loss: 0.0495 - acc: 0.9812 - val_loss: 0.7559 - val_acc: 0.7881\n",
      "Epoch 37/50\n",
      " - 2s - loss: 0.0410 - acc: 0.9878 - val_loss: 0.7680 - val_acc: 0.7947\n",
      "Epoch 38/50\n",
      " - 2s - loss: 0.0291 - acc: 0.9912 - val_loss: 0.7889 - val_acc: 0.7980\n",
      "Epoch 39/50\n",
      " - 2s - loss: 0.0342 - acc: 0.9878 - val_loss: 0.7816 - val_acc: 0.8179\n",
      "Epoch 40/50\n",
      " - 2s - loss: 0.0429 - acc: 0.9889 - val_loss: 0.7838 - val_acc: 0.7815\n",
      "Epoch 41/50\n",
      " - 2s - loss: 0.0416 - acc: 0.9889 - val_loss: 0.6956 - val_acc: 0.7980\n",
      "Epoch 42/50\n",
      " - 2s - loss: 0.0434 - acc: 0.9889 - val_loss: 0.6695 - val_acc: 0.8245\n",
      "Epoch 43/50\n",
      " - 2s - loss: 0.0370 - acc: 0.9889 - val_loss: 0.6278 - val_acc: 0.8179\n",
      "Epoch 44/50\n",
      " - 2s - loss: 0.0224 - acc: 0.9967 - val_loss: 0.6341 - val_acc: 0.8079\n",
      "Epoch 45/50\n",
      " - 2s - loss: 0.0289 - acc: 0.9945 - val_loss: 0.6983 - val_acc: 0.8113\n",
      "Epoch 46/50\n",
      " - 2s - loss: 0.0380 - acc: 0.9889 - val_loss: 0.7452 - val_acc: 0.8113\n",
      "Epoch 47/50\n",
      " - 2s - loss: 0.0369 - acc: 0.9900 - val_loss: 0.7629 - val_acc: 0.8013\n",
      "Epoch 48/50\n",
      " - 2s - loss: 0.0273 - acc: 0.9900 - val_loss: 0.7787 - val_acc: 0.8179\n",
      "Epoch 49/50\n",
      " - 2s - loss: 0.0386 - acc: 0.9900 - val_loss: 0.7534 - val_acc: 0.7980\n",
      "Epoch 50/50\n",
      " - 2s - loss: 0.0346 - acc: 0.9912 - val_loss: 0.7020 - val_acc: 0.7947\n"
     ]
    }
   ],
   "source": [
    "# Fit the model\n",
    "history=model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=50, batch_size=128, verbose=2)\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test, y_test, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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": [
    "import matplotlib.pyplot as plt\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",
    "# summarize history for loss\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": 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
}