SVM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import sklearn\n",
    "from sklearn.svm import SVC\n",
    "from sklearn import metrics\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "tickers = (\"fb\", \"aapl\", \"amzn\", \"nflx\", \"googl\")\n",
    "dfs = dict()\n",
    "models = dict()\n",
    "scores = dict()\n",
    "\n",
    "for i in tickers:\n",
    "    dfs[i] = pd.read_csv(i+\".csv\", index_col = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in tickers:\n",
    "\n",
    "    df = dfs[i]\n",
    "    \n",
    "    X = df.drop([\"signal\", \"Open\", \"High\", \"Low\", \"Close\"], axis = 1)\n",
    "    y = df['signal']\n",
    "    \n",
    "    X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X,y, train_size = 0.6)\n",
    "    \n",
    "    SVM_model = SVC(gamma = 'auto')\n",
    "    svc = SVM_model.fit(X_train, y_train)\n",
    "    models[i] = svc\n",
    "    scores[i] = SVM_model.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fb       0.503151\n",
       "aapl     0.487256\n",
       "amzn     0.505247\n",
       "nflx     0.501532\n",
       "googl    0.487298\n",
       "dtype: float64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(data = scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = pd.Series(svc.predict(X_test), index = y_test.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred.plot()\n",
    "plt.ylabel(\"Predicted signal\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred)\n",
    "auc = metrics.roc_auc_score(y_test, y_pred)\n",
    "plt.plot(fpr, tpr, label = \"data 1, auc=\"+str(auc))\n",
    "plt.legend(loc = 4)\n",
    "plt.title(\"auc for SVM\")\n",
    "plt.xlabel(\"false positive rate\")\n",
    "plt.ylabel(\"true positive rate\")\n",
    "plt.show()\n",
    "plt.savefig(\"auc_SVM.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Absolute Error:  0.5127020785219399\n",
      "Mean Squared Error:  0.5127020785219399\n",
      "Root Mean Squared Error:  0.7160321770157678\n"
     ]
    }
   ],
   "source": [
    "print(\"Mean Absolute Error: \",metrics.mean_absolute_error(y_test, y_pred))\n",
    "print(\"Mean Squared Error: \", metrics.mean_squared_error(y_test, y_pred))\n",
    "print(\"Root Mean Squared Error: \", np.sqrt(metrics.mean_squared_error(y_test, y_pred)))"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}