Regression - Stock forecast - Data, Preprocessing, fitting the model, forecastig and predicting.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import quandl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              Open    High     Low    Close      Volume  Ex-Dividend  \\\n",
      "Date                                                                   \n",
      "2004-08-19  100.01  104.06   95.96  100.335  44659000.0          0.0   \n",
      "2004-08-20  101.01  109.08  100.50  108.310  22834300.0          0.0   \n",
      "2004-08-23  110.76  113.48  109.05  109.400  18256100.0          0.0   \n",
      "2004-08-24  111.24  111.60  103.57  104.870  15247300.0          0.0   \n",
      "2004-08-25  104.76  108.00  103.88  106.000   9188600.0          0.0   \n",
      "\n",
      "            Split Ratio  Adj. Open  Adj. High   Adj. Low  Adj. Close  \\\n",
      "Date                                                                   \n",
      "2004-08-19          1.0  50.159839  52.191109  48.128568   50.322842   \n",
      "2004-08-20          1.0  50.661387  54.708881  50.405597   54.322689   \n",
      "2004-08-23          1.0  55.551482  56.915693  54.693835   54.869377   \n",
      "2004-08-24          1.0  55.792225  55.972783  51.945350   52.597363   \n",
      "2004-08-25          1.0  52.542193  54.167209  52.100830   53.164113   \n",
      "\n",
      "            Adj. Volume  \n",
      "Date                     \n",
      "2004-08-19   44659000.0  \n",
      "2004-08-20   22834300.0  \n",
      "2004-08-23   18256100.0  \n",
      "2004-08-24   15247300.0  \n",
      "2004-08-25    9188600.0  \n"
     ]
    }
   ],
   "source": [
    "df = quandl.get(\"WIKI/GOOGL\")\n",
    "\n",
    "print(df.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "too many columns, no point in having both regular and adjusted columns. .. we can remove regular columns and keep only adjusted columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[['Adj. Open',  'Adj. High',   'Adj. Low',  'Adj. Close', 'Adj. Volume']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Add High low percentage change .. may generate some insight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['HL_PCT'] = (df['Adj. High'] - df['Adj. Low']) / df['Adj. Low'] * 100.0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Close open percentage change"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['PCT_change'] = (df['Adj. Close'] - df['Adj. Open']) / df['Adj. Open'] * 100.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            Adj. Close    HL_PCT  PCT_change  Adj. Volume\n",
      "Date                                                     \n",
      "2004-08-19   50.322842  8.441017    0.324968   44659000.0\n",
      "2004-08-20   54.322689  8.537313    7.227007   22834300.0\n",
      "2004-08-23   54.869377  4.062357   -1.227880   18256100.0\n",
      "2004-08-24   52.597363  7.753210   -5.726357   15247300.0\n",
      "2004-08-25   53.164113  3.966115    1.183658    9188600.0\n"
     ]
    }
   ],
   "source": [
    "df = df[['Adj. Close', 'HL_PCT', 'PCT_change', 'Adj. Volume']]\n",
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import svm\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import preprocessing\n",
    "from sklearn.linear_model import LinearRegression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using numpy module to convert data to numpy array (scikit-learn uses it)\n",
    "preprocessing to clean and scale the data.. and train_text split is for testing stages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "forecast_col = 'Adj. Close'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fill Not a number (NaN) data with -99999-- using this as a outlier feature. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.fillna(value=-99999, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets assume we want to foreast 1% of the entire dataset length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "forecast_out = int(math.ceil(0.01 * len(df)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "add a new label .. shift our forecast column by desired number of periods.. Basically this is giving our actual prediction after 1% days /dataset period"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['label'] = df[forecast_col].shift(-forecast_out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            Adj. Close    HL_PCT  PCT_change  Adj. Volume      label\n",
      "Date                                                                \n",
      "2004-08-19   50.322842  8.441017    0.324968   44659000.0  69.639972\n",
      "2004-08-20   54.322689  8.537313    7.227007   22834300.0  69.078238\n",
      "2004-08-23   54.869377  4.062357   -1.227880   18256100.0  67.839414\n",
      "2004-08-24   52.597363  7.753210   -5.726357   15247300.0  68.912727\n",
      "2004-08-25   53.164113  3.966115    1.183658    9188600.0  70.668146\n"
     ]
    }
   ],
   "source": [
    "print(df.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is a good idea to drop anykind of NaN still there in the dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.dropna(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.array(df.drop(['label'], 1))\n",
    "y = np.array(df['label'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We should scale our feature in a range from -1 to 1. This will speed up the process and help with accuracy "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "X=preprocessing.scale(X)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "lets use SVM right now for classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = svm.SVR()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVR(C=1.0, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma='auto',\n",
       "  kernel='rbf', max_iter=-1, shrinking=True, tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "confidence = clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7874911789899935\n"
     ]
    }
   ],
   "source": [
    "print(confidence)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "accuracy of 78.7%\n",
    "Lets try another classifier \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = LinearRegression()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False)"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "confidence = clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9757146548770762\n"
     ]
    }
   ],
   "source": [
    "print(confidence)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Little better classifier .\n",
    "\n",
    "Now lets try with multiple treads. put n_jobs=-1 for using all the available threads. It will not make any differnce for such a small data, but it will make a huge difference if the data size is more than 20MB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = LinearRegression(n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=-1, normalize=False)"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "confidence = clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9757146548770762\n"
     ]
    }
   ],
   "source": [
    "print(confidence)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you want to try multiple kernel.kernel is like a transformation against the data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "linear 0.9741034974295733\n",
      "poly 0.6995347547432201\n",
      "rbf 0.7874911789899935\n",
      "sigmoid 0.8841608862870828\n"
     ]
    }
   ],
   "source": [
    "for k in ['linear','poly','rbf','sigmoid']:\n",
    "    clf = svm.SVR(kernel=k)\n",
    "    clf.fit(X_train, y_train)\n",
    "    confidence = clf.score(X_test, y_test)\n",
    "    print(k,confidence)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.array(df.drop(['label'], 1))\n",
    "X = preprocessing.scale(X)\n",
    "X_lately = X[-forecast_out:]\n",
    "X = X[:-forecast_out]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.dropna(inplace=True)\n",
    "y = np.array(df['label'])\n",
    "y = y[:-forecast_out]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = LinearRegression(n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=-1, normalize=False)"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "confidence = clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9736618913959262\n"
     ]
    }
   ],
   "source": [
    "print(confidence)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "forecast_set = clf.predict(X_lately)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1008.44964656 1048.23589205 1050.23203437 1050.6316576  1060.14523056\n",
      " 1060.51365991 1067.88190172 1060.9237701  1070.58527321 1065.32654767\n",
      " 1062.81413008 1059.4950288  1059.58768279 1054.71493837 1066.34595179\n",
      " 1053.40181369 1053.22989156 1068.42403981 1070.47254919 1075.0783405\n",
      " 1089.5148855  1080.2045872  1053.44526098 1053.65785953 1041.99955666\n",
      " 1028.5523286  1035.10700634 1049.40562261 1062.34426724 1067.49640641\n",
      " 1070.0882707  1066.19998515 1069.69226246 1075.37097138] 0.9736618913959262 34\n"
     ]
    }
   ],
   "source": [
    "print(forecast_set, confidence, forecast_out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "import datetime\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import style"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "import datetime object, matplotlib pyplot package for graphing and style create better looking graphs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "style.use('ggplot')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Forecast'] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "last_date = df.iloc[-1].name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "last_unix = last_date.timestamp()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "one_day = 86400\n",
    "next_unix = last_unix + one_day"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in forecast_set:\n",
    "    next_date = datetime.datetime.fromtimestamp(next_unix)\n",
    "    next_unix += 86400\n",
    "    df.loc[next_date] = [np.nan for _ in range(len(df.columns)-1)]+[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['Adj. Close'].plot()\n",
    "df['Forecast'].plot()\n",
    "plt.legend(loc=4)\n",
    "plt.xlabel('Date')\n",
    "plt.ylabel('Price')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### storing classifier with pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('linearregression.pickle','wb') as f:\n",
    "    pickle.dump(clf, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "pickle_in = open('linearregression.pickle','rb')\n",
    "clf = pickle.load(pickle_in)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "forecast_set = clf.predict(X_lately)\n",
    "df['Forecast'] = np.nan\n",
    "\n",
    "last_date = df.iloc[-1].name\n",
    "last_unix = last_date.timestamp()\n",
    "one_day = 86400\n",
    "next_unix = last_unix + one_day\n",
    "\n",
    "for i in forecast_set:\n",
    "    next_date = datetime.datetime.fromtimestamp(next_unix)\n",
    "    next_unix += 86400\n",
    "    df.loc[next_date] = [np.nan for _ in range(len(df.columns)-1)]+[i]\n",
    "df['Adj. Close'].plot()\n",
    "df['Forecast'].plot()\n",
    "plt.legend(loc=4)\n",
    "plt.xlabel('Date')\n",
    "plt.ylabel('Price')\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
}