Time Series LSTM VIX.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandas import read_csv\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "series = read_csv('vix_2011_2019.csv', header=0, parse_dates=[0], index_col=0,      \n",
    "                   squeeze=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                 Open       High    Low      Close  Volume  Adj Close\n",
      "Date                                                                 \n",
      "2011-02-11  16.530001  16.530001  15.55  15.690000       0  15.690000\n",
      "2011-02-14  16.070000  16.260000  15.22  15.950000       0  15.950000\n",
      "2011-02-15  16.299999  16.750000  16.27  16.370001       0  16.370001\n",
      "2011-02-16  16.309999  16.740000  15.84  16.719999       0  16.719999\n",
      "2011-02-17  17.010000  17.299999  15.88  16.590000       0  16.590000\n"
     ]
    }
   ],
   "source": [
    "print(series.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "series.drop(['Open', 'High', 'Low', 'Close', 'Volume'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "series.plot()\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "the adj close are quite erratic."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import time\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "from keras.layers.recurrent import LSTM\n",
    "from keras.models import Sequential"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def convertSeriesToMatrix(vectorSeries, sequence_length):\n",
    "    matrix=[]\n",
    "    for i in range(len(vectorSeries)-sequence_length+1):\n",
    "        matrix.append(vectorSeries[i:i+sequence_length])\n",
    "    return matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(2019)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "path_to_dataset = 'vix_2011_2019.csv'\n",
    "sequence_length = 20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# vector to store the time series\n",
    "vector_vix = []\n",
    "with open(path_to_dataset) as f:\n",
    "    next(f) # skip the header row\n",
    "    for line in f:\n",
    "        fields = line.split(',')\n",
    "        vector_vix.append(float(fields[6]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data shape:  (1992, 20)\n"
     ]
    }
   ],
   "source": [
    "# convert the vector to a 2D matrix\n",
    "matrix_vix = convertSeriesToMatrix(vector_vix, sequence_length)\n",
    "\n",
    "# shift all data by mean\n",
    "matrix_vix = np.array(matrix_vix)\n",
    "shifted_value = matrix_vix.mean()\n",
    "matrix_vix -= shifted_value\n",
    "print(\"Data shape: \", matrix_vix.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split dataset: 90% for training and 10% for testing\n",
    "train_row = int(round(0.9 * matrix_vix.shape[0]))\n",
    "train_set = matrix_vix[:train_row, :]\n",
    "\n",
    "# shuffle the training set (but do not shuffle the test set)\n",
    "np.random.shuffle(train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# the training set\n",
    "X_train = train_set[:, :-1]\n",
    "# the last column is the true value to compute the mean-squared-error loss\n",
    "y_train = train_set[:, -1] \n",
    "# the test set\n",
    "X_test = matrix_vix[train_row:, :-1]\n",
    "y_test = matrix_vix[train_row:, -1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# the input to LSTM layer needs to have the shape of (number of samples, the dimension of each element)\n",
    "X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 1))\n",
    "X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1793, 19, 1)\n",
      "(199, 19, 1)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape)\n",
    "print(X_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\SusanLi\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:3: UserWarning: The `input_dim` and `input_length` arguments in recurrent layers are deprecated. Use `input_shape` instead.\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n",
      "C:\\Users\\SusanLi\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:3: UserWarning: Update your `LSTM` call to the Keras 2 API: `LSTM(return_sequences=True, input_shape=(None, 1), units=50)`\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n",
      "C:\\Users\\SusanLi\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:6: UserWarning: Update your `LSTM` call to the Keras 2 API: `LSTM(return_sequences=False, units=100)`\n",
      "  \n",
      "C:\\Users\\SusanLi\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:10: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(activation=\"linear\", units=1)`\n",
      "  # Remove the CWD from sys.path while we load stuff.\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "# layer 1: LSTM\n",
    "model.add(LSTM( input_dim=1, output_dim=50, return_sequences=True))\n",
    "model.add(Dropout(0.2))\n",
    "# layer 2: LSTM\n",
    "model.add(LSTM(output_dim=100, return_sequences=False))\n",
    "model.add(Dropout(0.2))\n",
    "# layer 3: dense\n",
    "# linear activation: a(x) = x\n",
    "model.add(Dense(output_dim=1, activation='linear'))\n",
    "# compile the model\n",
    "model.compile(loss=\"mse\", optimizer=\"rmsprop\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1703 samples, validate on 90 samples\n",
      "Epoch 1/50\n",
      "1703/1703 [==============================] - 689s 404ms/step - loss: 26.8567 - val_loss: 23.8590\n",
      "Epoch 2/50\n",
      "1703/1703 [==============================] - 690s 405ms/step - loss: 16.9654 - val_loss: 17.0576\n",
      "Epoch 3/50\n",
      "1703/1703 [==============================] - 702s 412ms/step - loss: 13.0254 - val_loss: 13.1334\n",
      "Epoch 4/50\n",
      "1703/1703 [==============================] - 690s 405ms/step - loss: 10.7345 - val_loss: 10.5180\n",
      "Epoch 5/50\n",
      "1703/1703 [==============================] - 707s 415ms/step - loss: 9.2058 - val_loss: 9.2851\n",
      "Epoch 6/50\n",
      "1703/1703 [==============================] - 693s 407ms/step - loss: 8.2236 - val_loss: 8.6320\n",
      "Epoch 7/50\n",
      "1703/1703 [==============================] - 719s 422ms/step - loss: 7.6380 - val_loss: 8.0869\n",
      "Epoch 8/50\n",
      "1703/1703 [==============================] - 685s 402ms/step - loss: 7.5873 - val_loss: 7.5897\n",
      "Epoch 9/50\n",
      "1703/1703 [==============================] - 686s 403ms/step - loss: 6.5518 - val_loss: 7.3524\n",
      "Epoch 10/50\n",
      "1703/1703 [==============================] - 714s 419ms/step - loss: 6.2440 - val_loss: 6.6180\n",
      "Epoch 11/50\n",
      "1703/1703 [==============================] - 702s 413ms/step - loss: 6.0529 - val_loss: 6.6016\n",
      "Epoch 12/50\n",
      "1703/1703 [==============================] - 669s 393ms/step - loss: 5.8623 - val_loss: 6.3057\n",
      "Epoch 13/50\n",
      "1703/1703 [==============================] - 666s 391ms/step - loss: 5.3311 - val_loss: 5.7438\n",
      "Epoch 14/50\n",
      "1703/1703 [==============================] - 666s 391ms/step - loss: 5.1699 - val_loss: 5.5621\n",
      "Epoch 15/50\n",
      "1703/1703 [==============================] - 662s 389ms/step - loss: 4.9119 - val_loss: 5.1656\n",
      "Epoch 16/50\n",
      "1703/1703 [==============================] - 660s 388ms/step - loss: 4.6700 - val_loss: 5.1196\n",
      "Epoch 17/50\n",
      "1703/1703 [==============================] - 656s 385ms/step - loss: 4.6749 - val_loss: 4.3935\n",
      "Epoch 18/50\n",
      "1703/1703 [==============================] - 656s 385ms/step - loss: 4.3944 - val_loss: 4.9205\n",
      "Epoch 19/50\n",
      "1703/1703 [==============================] - 684s 402ms/step - loss: 4.2390 - val_loss: 4.0033\n",
      "Epoch 20/50\n",
      "1703/1703 [==============================] - 661s 388ms/step - loss: 3.9649 - val_loss: 3.9222\n",
      "Epoch 21/50\n",
      "1703/1703 [==============================] - 662s 389ms/step - loss: 3.7970 - val_loss: 3.4275\n",
      "Epoch 22/50\n",
      "1703/1703 [==============================] - 663s 389ms/step - loss: 4.3960 - val_loss: 3.6216\n",
      "Epoch 23/50\n",
      "1703/1703 [==============================] - 655s 385ms/step - loss: 3.6498 - val_loss: 3.2641\n",
      "Epoch 24/50\n",
      "1703/1703 [==============================] - 657s 386ms/step - loss: 3.6326 - val_loss: 3.1720\n",
      "Epoch 25/50\n",
      "1703/1703 [==============================] - 656s 385ms/step - loss: 3.4461 - val_loss: 3.2509\n",
      "Epoch 26/50\n",
      "1703/1703 [==============================] - 657s 386ms/step - loss: 3.6101 - val_loss: 3.5102\n",
      "Epoch 27/50\n",
      "1703/1703 [==============================] - 657s 386ms/step - loss: 3.4969 - val_loss: 3.5689\n",
      "Epoch 28/50\n",
      "1703/1703 [==============================] - 661s 388ms/step - loss: 3.2519 - val_loss: 3.1597\n",
      "Epoch 29/50\n",
      "1703/1703 [==============================] - 657s 386ms/step - loss: 3.3395 - val_loss: 3.1351\n",
      "Epoch 30/50\n",
      "1703/1703 [==============================] - 659s 387ms/step - loss: 3.3243 - val_loss: 3.1952\n",
      "Epoch 31/50\n",
      "1703/1703 [==============================] - 657s 386ms/step - loss: 3.0864 - val_loss: 2.8559\n",
      "Epoch 32/50\n",
      "1703/1703 [==============================] - 660s 387ms/step - loss: 3.1891 - val_loss: 3.9224\n",
      "Epoch 33/50\n",
      "1703/1703 [==============================] - 661s 388ms/step - loss: 3.3144 - val_loss: 2.7105\n",
      "Epoch 34/50\n",
      "1703/1703 [==============================] - 657s 386ms/step - loss: 3.0602 - val_loss: 2.7887\n",
      "Epoch 35/50\n",
      "1703/1703 [==============================] - 661s 388ms/step - loss: 3.1249 - val_loss: 3.1618\n",
      "Epoch 36/50\n",
      "1703/1703 [==============================] - 660s 387ms/step - loss: 3.2398 - val_loss: 2.7229\n",
      "Epoch 37/50\n",
      "1703/1703 [==============================] - 661s 388ms/step - loss: 3.0913 - val_loss: 2.7335\n",
      "Epoch 38/50\n",
      "1703/1703 [==============================] - 660s 388ms/step - loss: 3.0286 - val_loss: 2.7678\n",
      "Epoch 39/50\n",
      "1703/1703 [==============================] - 658s 386ms/step - loss: 2.9997 - val_loss: 3.1088\n",
      "Epoch 40/50\n",
      "1703/1703 [==============================] - 658s 386ms/step - loss: 3.0303 - val_loss: 2.6195\n",
      "Epoch 41/50\n",
      "1703/1703 [==============================] - 662s 389ms/step - loss: 2.9207 - val_loss: 2.7272\n",
      "Epoch 42/50\n",
      "1703/1703 [==============================] - 676s 397ms/step - loss: 3.0041 - val_loss: 2.6322\n",
      "Epoch 43/50\n",
      "1703/1703 [==============================] - 728s 427ms/step - loss: 2.7933 - val_loss: 2.3006\n",
      "Epoch 44/50\n",
      "1703/1703 [==============================] - 802s 471ms/step - loss: 2.9509 - val_loss: 2.4902\n",
      "Epoch 45/50\n",
      "1703/1703 [==============================] - 793s 465ms/step - loss: 2.8972 - val_loss: 2.5689\n",
      "Epoch 46/50\n",
      "1703/1703 [==============================] - 750s 440ms/step - loss: 2.8182 - val_loss: 2.4934\n",
      "Epoch 47/50\n",
      "1703/1703 [==============================] - 699s 410ms/step - loss: 2.7576 - val_loss: 2.4171\n",
      "Epoch 48/50\n",
      "1703/1703 [==============================] - 741s 435ms/step - loss: 2.6816 - val_loss: 2.3632\n",
      "Epoch 49/50\n",
      "1703/1703 [==============================] - 721s 423ms/step - loss: 2.7530 - val_loss: 2.7498\n",
      "Epoch 50/50\n",
      "1703/1703 [==============================] - 742s 436ms/step - loss: 2.7686 - val_loss: 2.2385\n",
      "199/199 [==============================] - 33s 168ms/step\n",
      "The mean squared error (MSE) on the test data set is 2.484 over 199 test samples.\n"
     ]
    }
   ],
   "source": [
    "model.fit(X_train, y_train, batch_size=512, epochs=50, validation_split=0.05, verbose=1)\n",
    "\n",
    "# evaluate the result\n",
    "test_mse = model.evaluate(X_test, y_test, verbose=1)\n",
    "print('The mean squared error (MSE) on the test data set is %.3f over %d test samples.' % (test_mse, len(y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get the predicted values\n",
    "predicted_values = model.predict(X_test)\n",
    "num_test_samples = len(predicted_values)\n",
    "predicted_values = np.reshape(predicted_values, (num_test_samples,1))\n",
    "\n",
    "# plot the results\n",
    "fig = plt.figure(figsize=(10,6))\n",
    "plt.plot(y_test + shifted_value)\n",
    "plt.plot(predicted_values + shifted_value)\n",
    "plt.xlabel('Date')\n",
    "plt.ylabel('VIX')\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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}