LSTM+neural+network.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "/Users/lukwingsan/anaconda3/lib/python3.6/importlib/_bootstrap.py:205: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6\n",
      "  return f(*args, **kwds)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time=1, Predicted=2.220940, Expected=2.960419\n",
      "Time=2, Predicted=3.246067, Expected=3.270605\n",
      "Time=3, Predicted=3.338441, Expected=3.028169\n",
      "Time=4, Predicted=2.860096, Expected=3.044829\n",
      "Time=5, Predicted=3.036892, Expected=2.764571\n",
      "Time=6, Predicted=2.608998, Expected=1.601082\n",
      "Time=7, Predicted=1.063832, Expected=1.874484\n",
      "Time=8, Predicted=2.087341, Expected=2.485244\n",
      "Time=9, Predicted=2.764305, Expected=2.767511\n",
      "Time=10, Predicted=2.864035, Expected=2.797886\n",
      "Time=11, Predicted=2.784121, Expected=3.342979\n",
      "Time=12, Predicted=3.578573, Expected=2.492586\n",
      "Time=13, Predicted=2.039095, Expected=1.884670\n",
      "Time=14, Predicted=1.644340, Expected=1.279494\n",
      "Time=15, Predicted=1.039755, Expected=1.287793\n",
      "Time=16, Predicted=1.352023, Expected=1.266283\n",
      "Time=17, Predicted=1.277772, Expected=1.515864\n",
      "Time=18, Predicted=1.651972, Expected=1.280262\n",
      "Time=19, Predicted=1.162138, Expected=2.953037\n",
      "Time=20, Predicted=3.750702, Expected=4.327243\n",
      "Time=21, Predicted=4.832002, Expected=4.297241\n",
      "Time=22, Predicted=4.146804, Expected=4.283669\n",
      "Time=23, Predicted=4.216656, Expected=4.511093\n",
      "Time=24, Predicted=4.562042, Expected=3.459267\n",
      "Time=25, Predicted=2.910822, Expected=1.859228\n",
      "Time=26, Predicted=1.159236, Expected=1.574420\n",
      "Time=27, Predicted=1.563675, Expected=1.563830\n",
      "Time=28, Predicted=1.614215, Expected=1.857848\n",
      "Time=29, Predicted=2.034378, Expected=1.234303\n",
      "Time=30, Predicted=0.948992, Expected=2.091823\n",
      "Time=31, Predicted=2.554431, Expected=2.725181\n",
      "Time=32, Predicted=2.967726, Expected=3.708423\n",
      "Time=33, Predicted=4.112734, Expected=4.402941\n",
      "Time=34, Predicted=4.625553, Expected=4.695037\n",
      "Time=35, Predicted=4.735129, Expected=4.331041\n",
      "Time=36, Predicted=4.087010, Expected=3.648534\n",
      "Time=37, Predicted=3.309304, Expected=3.609668\n",
      "Time=38, Predicted=3.613217, Expected=2.611553\n",
      "Time=39, Predicted=2.140396, Expected=2.592873\n",
      "Time=40, Predicted=2.661904, Expected=2.302856\n",
      "Time=41, Predicted=2.185855, Expected=2.941829\n",
      "Time=42, Predicted=3.280944, Expected=1.772882\n",
      "Time=43, Predicted=1.195425, Expected=1.783332\n",
      "Time=44, Predicted=1.881523, Expected=1.492260\n",
      "Time=45, Predicted=1.372117, Expected=1.280757\n",
      "Time=46, Predicted=1.223217, Expected=1.306708\n",
      "Time=47, Predicted=1.351284, Expected=1.537813\n",
      "Time=48, Predicted=1.658306, Expected=1.829515\n",
      "Time=49, Predicted=1.955876, Expected=1.838422\n",
      "Time=50, Predicted=1.818780, Expected=2.064698\n",
      "Time=51, Predicted=2.158788, Expected=1.441920\n",
      "Time=52, Predicted=1.122254, Expected=1.436810\n",
      "Time=53, Predicted=1.468101, Expected=0.917287\n",
      "Time=54, Predicted=0.672635, Expected=0.912771\n",
      "Time=55, Predicted=0.952161, Expected=1.857095\n",
      "Time=56, Predicted=2.304425, Expected=1.871738\n",
      "Time=57, Predicted=1.811145, Expected=2.473111\n",
      "Time=58, Predicted=2.736216, Expected=2.472577\n",
      "Time=59, Predicted=2.405089, Expected=2.245391\n",
      "Time=60, Predicted=2.111774, Expected=1.294068\n",
      "Time=61, Predicted=0.842462, Expected=1.277982\n",
      "Time=62, Predicted=1.330100, Expected=0.959039\n",
      "Time=63, Predicted=0.818050, Expected=1.167197\n",
      "Time=64, Predicted=1.299538, Expected=1.993934\n",
      "Time=65, Predicted=2.372975, Expected=2.281130\n",
      "Time=66, Predicted=2.353781, Expected=3.230958\n",
      "Time=67, Predicted=3.633180, Expected=2.723974\n",
      "Time=68, Predicted=2.389842, Expected=3.023397\n",
      "Time=69, Predicted=3.162100, Expected=2.415655\n",
      "Time=70, Predicted=2.081712, Expected=1.908645\n",
      "Time=71, Predicted=1.696522, Expected=0.958222\n",
      "Time=72, Predicted=0.546240, Expected=1.178095\n",
      "Time=73, Predicted=1.363715, Expected=1.189366\n",
      "Time=74, Predicted=1.201493, Expected=1.487332\n",
      "Time=75, Predicted=1.643767, Expected=1.487258\n",
      "Time=76, Predicted=1.473078, Expected=1.261509\n",
      "Time=77, Predicted=1.155599, Expected=1.565853\n",
      "Time=78, Predicted=1.720342, Expected=1.275460\n",
      "Time=79, Predicted=1.114627, Expected=0.986263\n",
      "Time=80, Predicted=0.866705, Expected=0.986046\n",
      "Time=81, Predicted=1.004522, Expected=1.839663\n",
      "Test RMSE: 0.586\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW4AAAD8CAYAAABXe05zAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x129652ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "from pandas import DataFrame\n",
    "from pandas import Series\n",
    "from pandas import concat\n",
    "from pandas import read_csv\n",
    "from pandas import datetime\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import LSTM\n",
    "from math import sqrt\n",
    "from matplotlib import pyplot\n",
    "import numpy\n",
    "\n",
    "# date-time parsing function for loading the dataset\n",
    "#def parser(x):\n",
    "#    return datetime.strptime('190'+x, '%f')\n",
    "\n",
    "# frame a sequence as a supervised learning problem\n",
    "def timeseries_to_supervised(data, lag=1):\n",
    "    df = DataFrame(data)\n",
    "    columns = [df.shift(i) for i in range(1, lag+1)]\n",
    "    columns.append(df)\n",
    "    df = concat(columns, axis=1)\n",
    "    df.fillna(0, inplace=True)\n",
    "    return df\n",
    "\n",
    "# create a differenced series\n",
    "def difference(dataset, interval=1):\n",
    "    diff = list()\n",
    "    for i in range(interval, len(dataset)):\n",
    "        value = dataset[i] - dataset[i - interval]\n",
    "        diff.append(value)\n",
    "    return Series(diff)\n",
    "\n",
    "# invert differenced value\n",
    "def inverse_difference(history, yhat, interval=1):\n",
    "    return yhat + history[-interval]\n",
    "\n",
    "# scale train and test data to [-1, 1]\n",
    "def scale(train, test):\n",
    "    # fit scaler\n",
    "    scaler = MinMaxScaler(feature_range=(-1, 1))\n",
    "    scaler = scaler.fit(train)\n",
    "    # transform train\n",
    "    train = train.reshape(train.shape[0],train.shape[1])\n",
    "    train_scaled = scaler.transform(train)\n",
    "    # transform test\n",
    "    test = test.reshape(test.shape[0],test.shape[1])\n",
    "    test_scaled = scaler.transform(test)\n",
    "    return scaler, train_scaled, test_scaled\n",
    "\n",
    "# inverse scaling for a forecasted value\n",
    "def invert_scale(scaler, X, value):\n",
    "    new_row = [x for x in X] + [value]\n",
    "    array = numpy.array(new_row)\n",
    "    array = array.reshape(1, len(array))\n",
    "    inverted = scaler.inverse_transform(array)\n",
    "    return inverted[0, -1]\n",
    "\n",
    "# fit an LSTM network to training data\n",
    "def fit_lstm(train, batch_size, nb_epoch, neurons):\n",
    "    X, y = train[:, 0:-1], train[:, -1]\n",
    "    X = X.reshape(X.shape[0], 1, X.shape[1])\n",
    "    model = Sequential()\n",
    "    model.add(LSTM(neurons, batch_input_shape=(batch_size, X.shape[1], X.shape[2]), stateful=True))\n",
    "    model.add(Dense(1))\n",
    "    model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "    for i in range(nb_epoch):\n",
    "        model.fit(X, y, epochs=1, batch_size=batch_size, verbose=0, shuffle=False)\n",
    "        model.reset_states()\n",
    "    return model\n",
    "\n",
    "# make a one-step forecast\n",
    "def forecast_lstm(model, batch_size, X):\n",
    "    X = X.reshape(1, 1, len(X))\n",
    "    yhat = model.predict(X, batch_size=batch_size)\n",
    "    return yhat[0,0]\n",
    "\n",
    "# load dataset\n",
    "series = read_csv('WorkbookFYP4.csv', header=0, parse_dates=[0], index_col=0, squeeze=True)\n",
    "\n",
    "# transform data to be stationary\n",
    "raw_values = series.values\n",
    "diff_values = difference(raw_values, 1)\n",
    "\n",
    "# transform data to be supervised learning\n",
    "supervised = timeseries_to_supervised(diff_values, 1)\n",
    "supervised_values = supervised.values\n",
    "\n",
    "# split data into train and test-sets\n",
    "train, test = supervised_values[0:-81], supervised_values[-81:]\n",
    "\n",
    "# transform the scale of the data\n",
    "scaler, train_scaled, test_scaled = scale(train, test)\n",
    "\n",
    "# fit the model\n",
    "lstm_model = fit_lstm(train_scaled, 1, 100, 4)\n",
    "# forecast the entire training dataset to build up state for forecasting\n",
    "train_reshaped = train_scaled[:, 0].reshape(len(train_scaled), 1, 1)\n",
    "lstm_model.predict(train_reshaped, batch_size=1)\n",
    "\n",
    "# walk-forward validation on the test data\n",
    "predictions = list()\n",
    "for i in range(len(test_scaled)):\n",
    "    # make one-step forecast\n",
    "    X, y = test_scaled[i, 0:-1], test_scaled[i, -1]\n",
    "    \n",
    "    \n",
    "    yhat = forecast_lstm(lstm_model, 1, X)\n",
    "\n",
    "\n",
    "   \n",
    "    # invert scaling\n",
    "    yhat = invert_scale(scaler, X, yhat)\n",
    "    # invert differencing\n",
    "    yhat = inverse_difference(raw_values, yhat, len(test_scaled)+1-i)\n",
    "    # store forecast\n",
    "    predictions.append(yhat)\n",
    "    expected = raw_values[len(train) + i + 1]\n",
    "    print('Time=%d, Predicted=%f, Expected=%f' % (i+1, yhat, expected))\n",
    "\n",
    "# report performance\n",
    "rmse = sqrt(mean_squared_error(raw_values[-81:], predictions))\n",
    "print('Test RMSE: %.3f' % rmse)\n",
    "# line plot of observed vs predicted\n",
    "pyplot.plot(raw_values[-81:])\n",
    "pyplot.plot(predictions)\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}