RNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# First step, import libraries.\n",
    "import numpy as np \n",
    "import pandas as pd \n",
    "from matplotlib import pyplot as plt\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Import the dataset and encode the date\n",
    "df = pd.read_csv('C:/Users/Divya Admin/Downloads/bitcoin-historical-data/bitstampUSD_1-min_data_2012-01-01_to_2018-01-08.csv')\n",
    "df['date'] = pd.to_datetime(df['Timestamp'],unit='s').dt.date\n",
    "group = df.groupby('date')\n",
    "Real_Price = group['Weighted_Price'].mean()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# split data\n",
    "prediction_days = 30\n",
    "df_train= Real_Price[:len(Real_Price)-prediction_days]\n",
    "df_test= Real_Price[len(Real_Price)-prediction_days:]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Data preprocess\n",
    "training_set = df_train.values\n",
    "training_set = np.reshape(training_set, (len(training_set), 1))\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "sc = MinMaxScaler()\n",
    "training_set = sc.fit_transform(training_set)\n",
    "X_train = training_set[0:len(training_set)-1]\n",
    "y_train = training_set[1:len(training_set)]\n",
    "X_train = np.reshape(X_train, (len(X_train), 1, 1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "2167/2167 [==============================] - 3s 2ms/step - loss: 0.0526\n",
      "Epoch 2/100\n",
      "2167/2167 [==============================] - 1s 629us/step - loss: 0.0079\n",
      "Epoch 3/100\n",
      "2167/2167 [==============================] - 1s 627us/step - loss: 0.0072\n",
      "Epoch 4/100\n",
      "2167/2167 [==============================] - 1s 614us/step - loss: 0.0065\n",
      "Epoch 5/100\n",
      "2167/2167 [==============================] - 1s 663us/step - loss: 0.0058\n",
      "Epoch 6/100\n",
      "2167/2167 [==============================] - 1s 623us/step - loss: 0.0051\n",
      "Epoch 7/100\n",
      "2167/2167 [==============================] - 1s 616us/step - loss: 0.0043\n",
      "Epoch 8/100\n",
      "2167/2167 [==============================] - 1s 687us/step - loss: 0.0035\n",
      "Epoch 9/100\n",
      "2167/2167 [==============================] - 1s 648us/step - loss: 0.0027\n",
      "Epoch 10/100\n",
      "2167/2167 [==============================] - 2s 706us/step - loss: 0.0019\n",
      "Epoch 11/100\n",
      "2167/2167 [==============================] - 1s 684us/step - loss: 0.0012\n",
      "Epoch 12/100\n",
      "2167/2167 [==============================] - 1s 651us/step - loss: 6.8662e-04\n",
      "Epoch 13/100\n",
      "2167/2167 [==============================] - 1s 637us/step - loss: 3.2557e-04\n",
      "Epoch 14/100\n",
      "2167/2167 [==============================] - 1s 640us/step - loss: 1.3073e-04\n",
      "Epoch 15/100\n",
      "2167/2167 [==============================] - 1s 650us/step - loss: 5.2268e-05\n",
      "Epoch 16/100\n",
      "2167/2167 [==============================] - 2s 706us/step - loss: 2.9944e-05\n",
      "Epoch 17/100\n",
      "2167/2167 [==============================] - 1s 637us/step - loss: 2.5487e-05\n",
      "Epoch 18/100\n",
      "2167/2167 [==============================] - 1s 655us/step - loss: 2.4658e-05\n",
      "Epoch 19/100\n",
      "2167/2167 [==============================] - 1s 642us/step - loss: 2.5162e-05\n",
      "Epoch 20/100\n",
      "2167/2167 [==============================] - 1s 645us/step - loss: 2.5354e-05\n",
      "Epoch 21/100\n",
      "2167/2167 [==============================] - 1s 640us/step - loss: 2.5979e-05\n",
      "Epoch 22/100\n",
      "2167/2167 [==============================] - 1s 649us/step - loss: 2.5769e-05\n",
      "Epoch 23/100\n",
      "2167/2167 [==============================] - 1s 661us/step - loss: 2.5897e-05\n",
      "Epoch 24/100\n",
      "2167/2167 [==============================] - 2s 703us/step - loss: 2.5748e-05\n",
      "Epoch 25/100\n",
      "2167/2167 [==============================] - 2s 722us/step - loss: 2.6589e-05\n",
      "Epoch 26/100\n",
      "2167/2167 [==============================] - 1s 666us/step - loss: 2.6223e-05\n",
      "Epoch 27/100\n",
      "2167/2167 [==============================] - 2s 693us/step - loss: 2.5877e-05\n",
      "Epoch 28/100\n",
      "2167/2167 [==============================] - 1s 689us/step - loss: 2.6824e-05\n",
      "Epoch 29/100\n",
      "2167/2167 [==============================] - 2s 746us/step - loss: 2.5462e-05\n",
      "Epoch 30/100\n",
      "2167/2167 [==============================] - 2s 761us/step - loss: 2.5195e-05\n",
      "Epoch 31/100\n",
      "2167/2167 [==============================] - 2s 704us/step - loss: 2.7215e-05\n",
      "Epoch 32/100\n",
      "2167/2167 [==============================] - 2s 735us/step - loss: 2.7045e-05\n",
      "Epoch 33/100\n",
      "2167/2167 [==============================] - 2s 710us/step - loss: 2.6034e-05\n",
      "Epoch 34/100\n",
      "2167/2167 [==============================] - 2s 713us/step - loss: 2.7192e-05\n",
      "Epoch 35/100\n",
      "2167/2167 [==============================] - 2s 709us/step - loss: 2.5829e-05\n",
      "Epoch 36/100\n",
      "2167/2167 [==============================] - 2s 711us/step - loss: 2.5318e-05\n",
      "Epoch 37/100\n",
      "2167/2167 [==============================] - 2s 714us/step - loss: 2.6124e-05\n",
      "Epoch 38/100\n",
      "2167/2167 [==============================] - 2s 698us/step - loss: 2.6300e-05\n",
      "Epoch 39/100\n",
      "2167/2167 [==============================] - 1s 692us/step - loss: 2.5918e-05\n",
      "Epoch 40/100\n",
      "2167/2167 [==============================] - 1s 650us/step - loss: 2.6686e-05\n",
      "Epoch 41/100\n",
      "2167/2167 [==============================] - 2s 723us/step - loss: 2.6625e-05\n",
      "Epoch 42/100\n",
      "2167/2167 [==============================] - 1s 679us/step - loss: 2.6075e-05\n",
      "Epoch 43/100\n",
      "2167/2167 [==============================] - 1s 684us/step - loss: 2.5146e-05\n",
      "Epoch 44/100\n",
      "2167/2167 [==============================] - 1s 671us/step - loss: 2.5417e-05\n",
      "Epoch 45/100\n",
      "2167/2167 [==============================] - 1s 668us/step - loss: 2.6874e-05\n",
      "Epoch 46/100\n",
      "2167/2167 [==============================] - 1s 669us/step - loss: 2.5397e-05\n",
      "Epoch 47/100\n",
      "2167/2167 [==============================] - 1s 672us/step - loss: 2.6773e-05\n",
      "Epoch 48/100\n",
      "2167/2167 [==============================] - 1s 679us/step - loss: 2.4583e-05\n",
      "Epoch 49/100\n",
      "2167/2167 [==============================] - 1s 655us/step - loss: 3.0355e-05\n",
      "Epoch 50/100\n",
      "2167/2167 [==============================] - 1s 677us/step - loss: 2.5741e-05\n",
      "Epoch 51/100\n",
      "2167/2167 [==============================] - 1s 682us/step - loss: 2.5796e-05\n",
      "Epoch 52/100\n",
      "2167/2167 [==============================] - 1s 676us/step - loss: 2.5821e-05\n",
      "Epoch 53/100\n",
      "2167/2167 [==============================] - 1s 688us/step - loss: 2.5372e-05\n",
      "Epoch 54/100\n",
      "2167/2167 [==============================] - 1s 681us/step - loss: 2.6101e-05\n",
      "Epoch 55/100\n",
      "2167/2167 [==============================] - 1s 671us/step - loss: 2.5789e-05\n",
      "Epoch 56/100\n",
      "2167/2167 [==============================] - 2s 705us/step - loss: 2.6296e-05\n",
      "Epoch 57/100\n",
      "2167/2167 [==============================] - 2s 709us/step - loss: 2.6038e-05\n",
      "Epoch 58/100\n",
      "2167/2167 [==============================] - 2s 714us/step - loss: 2.6288e-05\n",
      "Epoch 59/100\n",
      "2167/2167 [==============================] - 1s 679us/step - loss: 2.5443e-05\n",
      "Epoch 60/100\n",
      "2167/2167 [==============================] - 1s 683us/step - loss: 2.7584e-05\n",
      "Epoch 61/100\n",
      "2167/2167 [==============================] - 1s 690us/step - loss: 2.4442e-05\n",
      "Epoch 62/100\n",
      "2167/2167 [==============================] - 1s 692us/step - loss: 2.4870e-05\n",
      "Epoch 63/100\n",
      "2167/2167 [==============================] - 1s 673us/step - loss: 2.5692e-05\n",
      "Epoch 64/100\n",
      "2167/2167 [==============================] - 2s 694us/step - loss: 2.6710e-05\n",
      "Epoch 65/100\n",
      "2167/2167 [==============================] - 2s 699us/step - loss: 2.6243e-05\n",
      "Epoch 66/100\n",
      "2167/2167 [==============================] - 1s 679us/step - loss: 2.6975e-05\n",
      "Epoch 67/100\n",
      "2167/2167 [==============================] - 1s 679us/step - loss: 2.6401e-05\n",
      "Epoch 68/100\n",
      "2167/2167 [==============================] - 2s 697us/step - loss: 2.6428e-05\n",
      "Epoch 69/100\n",
      "2167/2167 [==============================] - 2s 734us/step - loss: 2.7555e-05\n",
      "Epoch 70/100\n",
      "2167/2167 [==============================] - 2s 704us/step - loss: 2.6244e-05\n",
      "Epoch 71/100\n",
      "2167/2167 [==============================] - 2s 728us/step - loss: 2.6417e-05\n",
      "Epoch 72/100\n",
      "2167/2167 [==============================] - 2s 746us/step - loss: 2.5854e-05\n",
      "Epoch 73/100\n",
      "2167/2167 [==============================] - 2s 765us/step - loss: 2.5695e-05\n",
      "Epoch 74/100\n",
      "2167/2167 [==============================] - 2s 821us/step - loss: 2.5299e-05\n",
      "Epoch 75/100\n",
      "2167/2167 [==============================] - 2s 779us/step - loss: 2.6621e-05\n",
      "Epoch 76/100\n",
      "2167/2167 [==============================] - 2s 768us/step - loss: 2.5002e-05\n",
      "Epoch 77/100\n",
      "2167/2167 [==============================] - 2s 720us/step - loss: 2.5447e-05\n",
      "Epoch 78/100\n",
      "2167/2167 [==============================] - 2s 705us/step - loss: 2.6411e-05\n",
      "Epoch 79/100\n",
      "2167/2167 [==============================] - 1s 690us/step - loss: 2.5933e-05\n",
      "Epoch 80/100\n",
      "2167/2167 [==============================] - 2s 697us/step - loss: 2.6021e-05\n",
      "Epoch 81/100\n",
      "2167/2167 [==============================] - 2s 717us/step - loss: 2.6625e-05\n",
      "Epoch 82/100\n",
      "2167/2167 [==============================] - 2s 729us/step - loss: 2.5742e-05\n",
      "Epoch 83/100\n",
      "2167/2167 [==============================] - 2s 717us/step - loss: 2.7094e-05\n",
      "Epoch 84/100\n",
      "2167/2167 [==============================] - 1s 689us/step - loss: 2.7183e-05\n",
      "Epoch 85/100\n",
      "2167/2167 [==============================] - 2s 740us/step - loss: 2.6067e-05\n",
      "Epoch 86/100\n",
      "2167/2167 [==============================] - 2s 744us/step - loss: 2.5172e-05\n",
      "Epoch 87/100\n",
      "2167/2167 [==============================] - 2s 705us/step - loss: 2.5656e-05\n",
      "Epoch 88/100\n",
      "2167/2167 [==============================] - 2s 785us/step - loss: 2.6008e-05\n",
      "Epoch 89/100\n",
      "2167/2167 [==============================] - 2s 728us/step - loss: 2.7356e-05\n",
      "Epoch 90/100\n",
      "2167/2167 [==============================] - 2s 714us/step - loss: 2.5653e-05\n",
      "Epoch 91/100\n",
      "2167/2167 [==============================] - 2s 809us/step - loss: 2.6847e-05\n",
      "Epoch 92/100\n",
      "2167/2167 [==============================] - 2s 754us/step - loss: 2.5668e-05\n",
      "Epoch 93/100\n",
      "2167/2167 [==============================] - 2s 734us/step - loss: 2.5847e-05\n",
      "Epoch 94/100\n",
      "2167/2167 [==============================] - 2s 703us/step - loss: 2.6106e-05\n",
      "Epoch 95/100\n",
      "2167/2167 [==============================] - 2s 709us/step - loss: 2.5400e-05\n",
      "Epoch 96/100\n",
      "2167/2167 [==============================] - 2s 694us/step - loss: 2.5863e-05\n",
      "Epoch 97/100\n",
      "2167/2167 [==============================] - 2s 706us/step - loss: 2.6001e-05\n",
      "Epoch 98/100\n",
      "2167/2167 [==============================] - 2s 716us/step - loss: 2.7674e-05\n",
      "Epoch 99/100\n",
      "2167/2167 [==============================] - 2s 717us/step - loss: 2.6087e-05\n",
      "Epoch 100/100\n",
      "2167/2167 [==============================] - 2s 714us/step - loss: 2.6107e-05\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0xaa38b4afd0>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Importing the Keras libraries and packages\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import LSTM\n",
    "\n",
    "# Initialising the RNN\n",
    "regressor = Sequential()\n",
    "\n",
    "# Adding the input layer and the LSTM layer\n",
    "regressor.add(LSTM(units = 4, activation = 'sigmoid', input_shape = (None, 1)))\n",
    "\n",
    "# Adding the output layer\n",
    "regressor.add(Dense(units = 1))\n",
    "\n",
    "# Compiling the RNN\n",
    "regressor.compile(optimizer = 'adam', loss = 'mean_squared_error')\n",
    "\n",
    "# Fitting the RNN to the Training set\n",
    "regressor.fit(X_train, y_train, batch_size = 5, epochs = 100)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Making the predictions\n",
    "test_set = df_test.values\n",
    "inputs = np.reshape(test_set, (len(test_set), 1))\n",
    "inputs = sc.transform(inputs)\n",
    "inputs = np.reshape(inputs, (len(inputs), 1, 1))\n",
    "predicted_BTC_price = regressor.predict(inputs)\n",
    "predicted_BTC_price = sc.inverse_transform(predicted_BTC_price)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Visualising the results\n",
    "plt.figure(figsize=(25,15), dpi=80, facecolor='w', edgecolor='k')\n",
    "ax = plt.gca()  \n",
    "plt.plot(test_set, color = 'red', label = 'Real BTC Price')\n",
    "plt.plot(predicted_BTC_price, color = 'blue', label = 'Predicted BTC Price')\n",
    "plt.title('BTC Price Prediction', fontsize=40)\n",
    "df_test = df_test.reset_index()\n",
    "x=df_test.index\n",
    "labels = df_test['date']\n",
    "plt.xticks(x, labels, rotation = 'vertical')\n",
    "for tick in ax.xaxis.get_major_ticks():\n",
    "    tick.label1.set_fontsize(18)\n",
    "for tick in ax.yaxis.get_major_ticks():\n",
    "    tick.label1.set_fontsize(18)\n",
    "plt.xlabel('Time', fontsize=40)\n",
    "plt.ylabel('BTC Price(USD)', fontsize=40)\n",
    "plt.legend(loc=2, prop={'size': 25})\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}