multi_step_hakim.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "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",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from numpy import array\n",
    "import pandas_datareader.data as web\n",
    " \n",
    "# date-time parsing function for loading the dataset\n",
    "def parser(x):\n",
    "    return datetime.strptime('190'+x, '%Y-%m')\n",
    " \n",
    "# convert time series into supervised learning problem\n",
    "def series_to_supervised(data, n_in=1, n_out=1, dropnan=True):\n",
    "    n_vars = 1 if type(data) is list else data.shape[1]\n",
    "    df = DataFrame(data)\n",
    "    cols, names = list(), list()\n",
    "    # input sequence (t-n, ... t-1)\n",
    "    for i in range(n_in, 0, -1):\n",
    "        cols.append(df.shift(i))\n",
    "        names += [('var%d(t-%d)' % (j+1, i)) for j in range(n_vars)]\n",
    "    # forecast sequence (t, t+1, ... t+n)\n",
    "    for i in range(0, n_out):\n",
    "        cols.append(df.shift(-i))\n",
    "        if i == 0:\n",
    "            names += [('var%d(t)' % (j+1)) for j in range(n_vars)]\n",
    "        else:\n",
    "            names += [('var%d(t+%d)' % (j+1, i)) for j in range(n_vars)]\n",
    "    # put it all together\n",
    "    agg = concat(cols, axis=1)\n",
    "    agg.columns = names\n",
    "    # drop rows with NaN values\n",
    "    if dropnan:\n",
    "        agg.dropna(inplace=True)\n",
    "    return agg\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",
    "# transform series into train and test sets for supervised learning\n",
    "def prepare_data(series, n_test, n_lag, n_seq):\n",
    "    # extract raw values\n",
    "    raw_values = series.values\n",
    "    # transform data to be stationary\n",
    "    diff_series = difference(raw_values, 1)\n",
    "    diff_values = diff_series.values\n",
    "    diff_values = diff_values.reshape(len(diff_values), 1)\n",
    "    # rescale values to -1, 1\n",
    "    scaler = MinMaxScaler(feature_range=(-1, 1))\n",
    "    scaled_values = scaler.fit_transform(diff_values)\n",
    "    scaled_values = scaled_values.reshape(len(scaled_values), 1)\n",
    "    # transform into supervised learning problem X, y\n",
    "    supervised = series_to_supervised(scaled_values, n_lag, n_seq)\n",
    "    supervised_values = supervised.values\n",
    "    # split into train and test sets\n",
    "    train, test = supervised_values[0:-n_test], supervised_values[-n_test:]\n",
    "    return scaler, train, test\n",
    " \n",
    "# fit an LSTM network to training data\n",
    "def fit_lstm(train, n_lag, n_seq, n_batch, nb_epoch, n_neurons):\n",
    "    # reshape training into [samples, timesteps, features]\n",
    "    X, y = train[:, 0:n_lag], train[:, n_lag:]\n",
    "    X = X.reshape(X.shape[0], 1, X.shape[1])\n",
    "    # design network\n",
    "    model = Sequential()\n",
    "    model.add(LSTM(n_neurons, batch_input_shape=(n_batch, X.shape[1], X.shape[2]), stateful=True))\n",
    "    model.add(Dense(y.shape[1]))\n",
    "    model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "    # fit network\n",
    "    for i in range(nb_epoch):\n",
    "        print(\"Epoch: \" + str(i+1) + \"/\" + str(nb_epoch))\n",
    "        model.fit(X, y, epochs=1, batch_size=n_batch, verbose=0, shuffle=False)\n",
    "        model.reset_states()\n",
    "    return model\n",
    " \n",
    "# make one forecast with an LSTM,\n",
    "def forecast_lstm(model, X, n_batch):\n",
    "    # reshape input pattern to [samples, timesteps, features]\n",
    "    X = X.reshape(1, 1, len(X))\n",
    "    # make forecast\n",
    "    forecast = model.predict(X, batch_size=n_batch)\n",
    "    # convert to array\n",
    "    return [x for x in forecast[0, :]]\n",
    " \n",
    "# evaluate the persistence model\n",
    "def make_forecasts(model, n_batch, train, test, n_lag, n_seq):\n",
    "    forecasts = list()\n",
    "    for i in range(len(test)):\n",
    "        X, y = test[i, 0:n_lag], test[i, n_lag:]\n",
    "        # make forecast\n",
    "        forecast = forecast_lstm(model, X, n_batch)\n",
    "        # store the forecast\n",
    "        forecasts.append(forecast)\n",
    "    return forecasts\n",
    " \n",
    "# invert differenced forecast\n",
    "def inverse_difference(last_ob, forecast):\n",
    "    # invert first forecast\n",
    "    inverted = list()\n",
    "    inverted.append(forecast[0] + last_ob)\n",
    "    # propagate difference forecast using inverted first value\n",
    "    for i in range(1, len(forecast)):\n",
    "        inverted.append(forecast[i] + inverted[i-1])\n",
    "    return inverted\n",
    " \n",
    "# inverse data transform on forecasts\n",
    "def inverse_transform(series, forecasts, scaler, n_test):\n",
    "    inverted = list()\n",
    "    for i in range(len(forecasts)):\n",
    "        # create array from forecast\n",
    "        forecast = array(forecasts[i])\n",
    "        forecast = forecast.reshape(1, len(forecast))\n",
    "        # invert scaling\"Month\",\"Sales of shampoo over a three year period\"\n",
    "        inv_scale = scaler.inverse_transform(forecast)\n",
    "        inv_scale = inv_scale[0, :]\n",
    "        # invert differencing\n",
    "        index = len(series) - n_test + i - 1\n",
    "        last_ob = series.values[index]\n",
    "        inv_diff = inverse_difference(last_ob, inv_scale)\n",
    "        # store\n",
    "        inverted.append(inv_diff)\n",
    "    return inverted\n",
    " \n",
    "# evaluate the RMSE for each forecast time step\n",
    "def evaluate_forecasts(test, forecasts, n_lag, n_seq):\n",
    "    for i in range(n_seq):\n",
    "        actual = [row[i] for row in test]\n",
    "        predicted = [forecast[i] for forecast in forecasts]\n",
    "        rmse = sqrt(mean_squared_error(actual, predicted))\n",
    "        print('t+%d RMSE: %f' % ((i+1), rmse))\n",
    " \n",
    "# plot the forecasts in the context of the original dataset\n",
    "def plot_forecasts(series, forecasts, n_test):\n",
    "    # plot the entire dataset in blue\n",
    "    plt.plot(series.values)\n",
    "    # plot the forecasts in red\n",
    "    for i in range(len(forecasts)):\n",
    "        off_s = len(series) - n_test + i - 1\n",
    "        off_e = off_s + len(forecasts[i]) + 1\n",
    "        xaxis = [x for x in range(off_s, off_e)]\n",
    "        yaxis = [series.values[off_s]] + forecasts[i]\n",
    "        plt.plot(xaxis, yaxis, color='red')\n",
    "    # show the plot\n",
    "    plt.show()\n",
    " \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "...utils/validation.py:444: DataConversionWarning: Data with input dtype object was converted to float64 by MinMaxScaler.\n",
      "  warnings.warn(msg, DataConversionWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 0/500\n",
      "Epoch: 1/500\n",
      "Epoch: 2/500\n",
      "Epoch: 3/500\n",
      "Epoch: 4/500\n",
      "Epoch: 5/500\n",
      "Epoch: 6/500\n",
      "Epoch: 7/500\n",
      "Epoch: 8/500\n",
      "Epoch: 9/500\n",
      "Epoch: 10/500\n",
      "Epoch: 11/500\n",
      "Epoch: 12/500\n",
      "Epoch: 13/500\n",
      "Epoch: 14/500\n",
      "Epoch: 15/500\n",
      "Epoch: 16/500\n",
      "Epoch: 17/500\n",
      "Epoch: 18/500\n",
      "Epoch: 19/500\n",
      "Epoch: 20/500\n",
      "Epoch: 21/500\n",
      "Epoch: 22/500\n",
      "Epoch: 23/500\n",
      "Epoch: 24/500\n",
      "Epoch: 25/500\n",
      "Epoch: 26/500\n",
      "Epoch: 27/500\n",
      "Epoch: 28/500\n",
      "Epoch: 29/500\n",
      "Epoch: 30/500\n",
      "Epoch: 31/500\n",
      "Epoch: 32/500\n",
      "Epoch: 33/500\n",
      "Epoch: 34/500\n",
      "Epoch: 35/500\n",
      "Epoch: 36/500\n",
      "Epoch: 37/500\n",
      "Epoch: 38/500\n",
      "Epoch: 39/500\n",
      "Epoch: 40/500\n",
      "Epoch: 41/500\n",
      "Epoch: 42/500\n",
      "Epoch: 43/500\n",
      "Epoch: 44/500\n",
      "Epoch: 45/500\n",
      "Epoch: 46/500\n",
      "Epoch: 47/500\n",
      "Epoch: 48/500\n",
      "Epoch: 49/500\n",
      "Epoch: 50/500\n",
      "Epoch: 51/500\n",
      "Epoch: 52/500\n",
      "Epoch: 53/500\n",
      "Epoch: 54/500\n",
      "Epoch: 55/500\n",
      "Epoch: 56/500\n",
      "Epoch: 57/500\n",
      "Epoch: 58/500\n",
      "Epoch: 59/500\n",
      "Epoch: 60/500\n",
      "Epoch: 61/500\n",
      "Epoch: 62/500\n",
      "Epoch: 63/500\n",
      "Epoch: 64/500\n",
      "Epoch: 65/500\n",
      "Epoch: 66/500\n",
      "Epoch: 67/500\n",
      "Epoch: 68/500\n",
      "Epoch: 69/500\n",
      "Epoch: 70/500\n",
      "Epoch: 71/500\n",
      "Epoch: 72/500\n",
      "Epoch: 73/500\n",
      "Epoch: 74/500\n",
      "Epoch: 75/500\n",
      "Epoch: 76/500\n",
      "Epoch: 77/500\n",
      "Epoch: 78/500\n",
      "Epoch: 79/500\n",
      "Epoch: 80/500\n",
      "Epoch: 81/500\n",
      "Epoch: 82/500\n",
      "Epoch: 83/500\n",
      "Epoch: 84/500\n",
      "Epoch: 85/500\n",
      "Epoch: 86/500\n",
      "Epoch: 87/500\n",
      "Epoch: 88/500\n",
      "Epoch: 89/500\n",
      "Epoch: 90/500\n",
      "Epoch: 91/500\n",
      "Epoch: 92/500\n",
      "Epoch: 93/500\n",
      "Epoch: 94/500\n",
      "Epoch: 95/500\n",
      "Epoch: 96/500\n",
      "Epoch: 97/500\n",
      "Epoch: 98/500\n",
      "Epoch: 99/500\n",
      "Epoch: 100/500\n",
      "Epoch: 101/500\n",
      "Epoch: 102/500\n",
      "Epoch: 103/500\n",
      "Epoch: 104/500\n",
      "Epoch: 105/500\n",
      "Epoch: 106/500\n",
      "Epoch: 107/500\n",
      "Epoch: 108/500\n",
      "Epoch: 109/500\n",
      "Epoch: 110/500\n",
      "Epoch: 111/500\n",
      "Epoch: 112/500\n",
      "Epoch: 113/500\n",
      "Epoch: 114/500\n",
      "Epoch: 115/500\n",
      "Epoch: 116/500\n",
      "Epoch: 117/500\n",
      "Epoch: 118/500\n",
      "Epoch: 119/500\n",
      "Epoch: 120/500\n",
      "Epoch: 121/500\n",
      "Epoch: 122/500\n",
      "Epoch: 123/500\n",
      "Epoch: 124/500\n",
      "Epoch: 125/500\n",
      "Epoch: 126/500\n",
      "Epoch: 127/500\n",
      "Epoch: 128/500\n",
      "Epoch: 129/500\n",
      "Epoch: 130/500\n",
      "Epoch: 131/500\n",
      "Epoch: 132/500\n",
      "Epoch: 133/500\n",
      "Epoch: 134/500\n",
      "Epoch: 135/500\n",
      "Epoch: 136/500\n",
      "Epoch: 137/500\n",
      "Epoch: 138/500\n",
      "Epoch: 139/500\n",
      "Epoch: 140/500\n",
      "Epoch: 141/500\n",
      "Epoch: 142/500\n",
      "Epoch: 143/500\n",
      "Epoch: 144/500\n",
      "Epoch: 145/500\n",
      "Epoch: 146/500\n",
      "Epoch: 147/500\n",
      "Epoch: 148/500\n",
      "Epoch: 149/500\n",
      "Epoch: 150/500\n",
      "Epoch: 151/500\n",
      "Epoch: 152/500\n",
      "Epoch: 153/500\n",
      "Epoch: 154/500\n",
      "Epoch: 155/500\n",
      "Epoch: 156/500\n",
      "Epoch: 157/500\n",
      "Epoch: 158/500\n",
      "Epoch: 159/500\n",
      "Epoch: 160/500\n",
      "Epoch: 161/500\n",
      "Epoch: 162/500\n",
      "Epoch: 163/500\n",
      "Epoch: 164/500\n",
      "Epoch: 165/500\n",
      "Epoch: 166/500\n",
      "Epoch: 167/500\n",
      "Epoch: 168/500\n",
      "Epoch: 169/500\n",
      "Epoch: 170/500\n",
      "Epoch: 171/500\n",
      "Epoch: 172/500\n",
      "Epoch: 173/500\n",
      "Epoch: 174/500\n",
      "Epoch: 175/500\n",
      "Epoch: 176/500\n",
      "Epoch: 177/500\n",
      "Epoch: 178/500\n",
      "Epoch: 179/500\n",
      "Epoch: 180/500\n",
      "Epoch: 181/500\n",
      "Epoch: 182/500\n",
      "Epoch: 183/500\n",
      "Epoch: 184/500\n",
      "Epoch: 185/500\n",
      "Epoch: 186/500\n",
      "Epoch: 187/500\n",
      "Epoch: 188/500\n",
      "Epoch: 189/500\n",
      "Epoch: 190/500\n",
      "Epoch: 191/500\n",
      "Epoch: 192/500\n",
      "Epoch: 193/500\n",
      "Epoch: 194/500\n",
      "Epoch: 195/500\n",
      "Epoch: 196/500\n",
      "Epoch: 197/500\n",
      "Epoch: 198/500\n",
      "Epoch: 199/500\n",
      "Epoch: 200/500\n",
      "Epoch: 201/500\n",
      "Epoch: 202/500\n",
      "Epoch: 203/500\n",
      "Epoch: 204/500\n",
      "Epoch: 205/500\n",
      "Epoch: 206/500\n",
      "Epoch: 207/500\n",
      "Epoch: 208/500\n",
      "Epoch: 209/500\n",
      "Epoch: 210/500\n",
      "Epoch: 211/500\n",
      "Epoch: 212/500\n",
      "Epoch: 213/500\n",
      "Epoch: 214/500\n",
      "Epoch: 215/500\n",
      "Epoch: 216/500\n",
      "Epoch: 217/500\n",
      "Epoch: 218/500\n",
      "Epoch: 219/500\n",
      "Epoch: 220/500\n",
      "Epoch: 221/500\n",
      "Epoch: 222/500\n",
      "Epoch: 223/500\n",
      "Epoch: 224/500\n",
      "Epoch: 225/500\n",
      "Epoch: 226/500\n",
      "Epoch: 227/500\n",
      "Epoch: 228/500\n",
      "Epoch: 229/500\n",
      "Epoch: 230/500\n",
      "Epoch: 231/500\n",
      "Epoch: 232/500\n",
      "Epoch: 233/500\n",
      "Epoch: 234/500\n",
      "Epoch: 235/500\n",
      "Epoch: 236/500\n",
      "Epoch: 237/500\n",
      "Epoch: 238/500\n",
      "Epoch: 239/500\n",
      "Epoch: 240/500\n",
      "Epoch: 241/500\n",
      "Epoch: 242/500\n",
      "Epoch: 243/500\n",
      "Epoch: 244/500\n",
      "Epoch: 245/500\n",
      "Epoch: 246/500\n",
      "Epoch: 247/500\n",
      "Epoch: 248/500\n",
      "Epoch: 249/500\n",
      "Epoch: 250/500\n",
      "Epoch: 251/500\n",
      "Epoch: 252/500\n",
      "Epoch: 253/500\n",
      "Epoch: 254/500\n",
      "Epoch: 255/500\n",
      "Epoch: 256/500\n",
      "Epoch: 257/500\n",
      "Epoch: 258/500\n",
      "Epoch: 259/500\n",
      "Epoch: 260/500\n",
      "Epoch: 261/500\n",
      "Epoch: 262/500\n",
      "Epoch: 263/500\n",
      "Epoch: 264/500\n",
      "Epoch: 265/500\n",
      "Epoch: 266/500\n",
      "Epoch: 267/500\n",
      "Epoch: 268/500\n",
      "Epoch: 269/500\n",
      "Epoch: 270/500\n",
      "Epoch: 271/500\n",
      "Epoch: 272/500\n",
      "Epoch: 273/500\n",
      "Epoch: 274/500\n",
      "Epoch: 275/500\n",
      "Epoch: 276/500\n",
      "Epoch: 277/500\n",
      "Epoch: 278/500\n",
      "Epoch: 279/500\n",
      "Epoch: 280/500\n",
      "Epoch: 281/500\n",
      "Epoch: 282/500\n",
      "Epoch: 283/500\n",
      "Epoch: 284/500\n",
      "Epoch: 285/500\n",
      "Epoch: 286/500\n",
      "Epoch: 287/500\n",
      "Epoch: 288/500\n",
      "Epoch: 289/500\n",
      "Epoch: 290/500\n",
      "Epoch: 291/500\n",
      "Epoch: 292/500\n",
      "Epoch: 293/500\n",
      "Epoch: 294/500\n",
      "Epoch: 295/500\n",
      "Epoch: 296/500\n",
      "Epoch: 297/500\n",
      "Epoch: 298/500\n",
      "Epoch: 299/500\n",
      "Epoch: 300/500\n",
      "Epoch: 301/500\n",
      "Epoch: 302/500\n",
      "Epoch: 303/500\n",
      "Epoch: 304/500\n",
      "Epoch: 305/500\n",
      "Epoch: 306/500\n",
      "Epoch: 307/500\n",
      "Epoch: 308/500\n",
      "Epoch: 309/500\n",
      "Epoch: 310/500\n",
      "Epoch: 311/500\n",
      "Epoch: 312/500\n",
      "Epoch: 313/500\n",
      "Epoch: 314/500\n",
      "Epoch: 315/500\n",
      "Epoch: 316/500\n",
      "Epoch: 317/500\n",
      "Epoch: 318/500\n",
      "Epoch: 319/500\n",
      "Epoch: 320/500\n",
      "Epoch: 321/500\n",
      "Epoch: 322/500\n",
      "Epoch: 323/500\n",
      "Epoch: 324/500\n",
      "Epoch: 325/500\n",
      "Epoch: 326/500\n",
      "Epoch: 327/500\n",
      "Epoch: 328/500\n",
      "Epoch: 329/500\n",
      "Epoch: 330/500\n",
      "Epoch: 331/500\n",
      "Epoch: 332/500\n",
      "Epoch: 333/500\n",
      "Epoch: 334/500\n",
      "Epoch: 335/500\n",
      "Epoch: 336/500\n",
      "Epoch: 337/500\n",
      "Epoch: 338/500\n",
      "Epoch: 339/500\n",
      "Epoch: 340/500\n",
      "Epoch: 341/500\n",
      "Epoch: 342/500\n",
      "Epoch: 343/500\n",
      "Epoch: 344/500\n",
      "Epoch: 345/500\n",
      "Epoch: 346/500\n",
      "Epoch: 347/500\n",
      "Epoch: 348/500\n",
      "Epoch: 349/500\n",
      "Epoch: 350/500\n",
      "Epoch: 351/500\n",
      "Epoch: 352/500\n",
      "Epoch: 353/500\n",
      "Epoch: 354/500\n",
      "Epoch: 355/500\n",
      "Epoch: 356/500\n",
      "Epoch: 357/500\n",
      "Epoch: 358/500\n",
      "Epoch: 359/500\n",
      "Epoch: 360/500\n",
      "Epoch: 361/500\n",
      "Epoch: 362/500\n",
      "Epoch: 363/500\n",
      "Epoch: 364/500\n",
      "Epoch: 365/500\n",
      "Epoch: 366/500\n",
      "Epoch: 367/500\n",
      "Epoch: 368/500\n",
      "Epoch: 369/500\n",
      "Epoch: 370/500\n",
      "Epoch: 371/500\n",
      "Epoch: 372/500\n",
      "Epoch: 373/500\n",
      "Epoch: 374/500\n",
      "Epoch: 375/500\n",
      "Epoch: 376/500\n",
      "Epoch: 377/500\n",
      "Epoch: 378/500\n",
      "Epoch: 379/500\n",
      "Epoch: 380/500\n",
      "Epoch: 381/500\n",
      "Epoch: 382/500\n",
      "Epoch: 383/500\n",
      "Epoch: 384/500\n",
      "Epoch: 385/500\n",
      "Epoch: 386/500\n",
      "Epoch: 387/500\n",
      "Epoch: 388/500\n",
      "Epoch: 389/500\n",
      "Epoch: 390/500\n",
      "Epoch: 391/500\n",
      "Epoch: 392/500\n",
      "Epoch: 393/500\n",
      "Epoch: 394/500\n",
      "Epoch: 395/500\n",
      "Epoch: 396/500\n",
      "Epoch: 397/500\n",
      "Epoch: 398/500\n",
      "Epoch: 399/500\n",
      "Epoch: 400/500\n",
      "Epoch: 401/500\n",
      "Epoch: 402/500\n",
      "Epoch: 403/500\n",
      "Epoch: 404/500\n",
      "Epoch: 405/500\n",
      "Epoch: 406/500\n",
      "Epoch: 407/500\n",
      "Epoch: 408/500\n",
      "Epoch: 409/500\n",
      "Epoch: 410/500\n",
      "Epoch: 411/500\n",
      "Epoch: 412/500\n",
      "Epoch: 413/500\n",
      "Epoch: 414/500\n",
      "Epoch: 415/500\n",
      "Epoch: 416/500\n",
      "Epoch: 417/500\n",
      "Epoch: 418/500\n",
      "Epoch: 419/500\n",
      "Epoch: 420/500\n",
      "Epoch: 421/500\n",
      "Epoch: 422/500\n",
      "Epoch: 423/500\n",
      "Epoch: 424/500\n",
      "Epoch: 425/500\n",
      "Epoch: 426/500\n",
      "Epoch: 427/500\n",
      "Epoch: 428/500\n",
      "Epoch: 429/500\n",
      "Epoch: 430/500\n",
      "Epoch: 431/500\n",
      "Epoch: 432/500\n",
      "Epoch: 433/500\n",
      "Epoch: 434/500\n",
      "Epoch: 435/500\n",
      "Epoch: 436/500\n",
      "Epoch: 437/500\n",
      "Epoch: 438/500\n",
      "Epoch: 439/500\n",
      "Epoch: 440/500\n",
      "Epoch: 441/500\n",
      "Epoch: 442/500\n",
      "Epoch: 443/500\n",
      "Epoch: 444/500\n",
      "Epoch: 445/500\n",
      "Epoch: 446/500\n",
      "Epoch: 447/500\n",
      "Epoch: 448/500\n",
      "Epoch: 449/500\n",
      "Epoch: 450/500\n",
      "Epoch: 451/500\n",
      "Epoch: 452/500\n",
      "Epoch: 453/500\n",
      "Epoch: 454/500\n",
      "Epoch: 455/500\n",
      "Epoch: 456/500\n",
      "Epoch: 457/500\n",
      "Epoch: 458/500\n",
      "Epoch: 459/500\n",
      "Epoch: 460/500\n",
      "Epoch: 461/500\n",
      "Epoch: 462/500\n",
      "Epoch: 463/500\n",
      "Epoch: 464/500\n",
      "Epoch: 465/500\n",
      "Epoch: 466/500\n",
      "Epoch: 467/500\n",
      "Epoch: 468/500\n",
      "Epoch: 469/500\n",
      "Epoch: 470/500\n",
      "Epoch: 471/500\n",
      "Epoch: 472/500\n",
      "Epoch: 473/500\n",
      "Epoch: 474/500\n",
      "Epoch: 475/500\n",
      "Epoch: 476/500\n",
      "Epoch: 477/500\n",
      "Epoch: 478/500\n",
      "Epoch: 479/500\n",
      "Epoch: 480/500\n",
      "Epoch: 481/500\n",
      "Epoch: 482/500\n",
      "Epoch: 483/500\n",
      "Epoch: 484/500\n",
      "Epoch: 485/500\n",
      "Epoch: 486/500\n",
      "Epoch: 487/500\n",
      "Epoch: 488/500\n",
      "Epoch: 489/500\n",
      "Epoch: 490/500\n",
      "Epoch: 491/500\n",
      "Epoch: 492/500\n",
      "Epoch: 493/500\n",
      "Epoch: 494/500\n",
      "Epoch: 495/500\n",
      "Epoch: 496/500\n",
      "Epoch: 497/500\n",
      "Epoch: 498/500\n",
      "Epoch: 499/500\n",
      "t+1 RMSE: 3.376584\n",
      "t+2 RMSE: 6.055488\n",
      "t+3 RMSE: 7.818268\n",
      "t+4 RMSE: 8.251823\n",
      "t+5 RMSE: 6.587765\n",
      "t+6 RMSE: 4.539264\n",
      "t+7 RMSE: 2.717651\n",
      "t+8 RMSE: 1.909380\n",
      "t+9 RMSE: 1.522130\n",
      "t+10 RMSE: 2.735811\n"
     ]
    }
   ],
   "source": [
    "# Plotsize\n",
    "plt.rcParams['figure.figsize'] = [18,12]\n",
    "\n",
    "# load dataset\n",
    "#series = read_csv('shampoo-sales.csv', header=0, parse_dates=[0], index_col=0, squeeze=True, date_parser=parser)\n",
    "\n",
    "series = web.DataReader('^GSPC','yahoo')['Open'].to_frame()\n",
    "\n",
    "# configure\n",
    "n_lag = 1\n",
    "n_seq = 10 #How far ahead\n",
    "n_test = 3 #Nr tests\n",
    "n_epochs = 500 #Training time\n",
    "n_batch = 1\n",
    "n_neurons = 1\n",
    "time_shift = n_test + n_seq - 1\n",
    "\n",
    "# prepare data\n",
    "scaler, train, test = prepare_data(series, n_test, n_lag, n_seq)\n",
    "# fit model\n",
    "model = fit_lstm(train, n_lag, n_seq, n_batch, n_epochs, n_neurons)\n",
    "# make forecasts\n",
    "forecasts = make_forecasts(model, n_batch, train, test, n_lag, n_seq)\n",
    "# inverse transform forecasts and test\n",
    "forecasts = inverse_transform(series, forecasts, scaler, time_shift)\n",
    "actual = [row[n_lag:] for row in test]\n",
    "actual = inverse_transform(series, actual, scaler, time_shift)\n",
    "# evaluate forecasts\n",
    "evaluate_forecasts(actual, forecasts, n_lag, n_seq)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABB4AAAKvCAYAAADN4AzZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8400bdb0b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot forecasts\n",
    "plot_forecasts(series.tail(30), forecasts, time_shift)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tensorflow",
   "language": "python",
   "name": "tensorflow"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.2"
  },
  "nbTranslate": {
   "displayLangs": [
    "en",
    "nl"
   ],
   "hotkey": "alt-t",
   "langInMainMenu": true,
   "sourceLang": "nl",
   "targetLang": "en",
   "useGoogleTranslate": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}