PCA + SVM Stacking.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import sqlite3\n",
    "from sklearn.svm import LinearSVC\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import style\n",
    "from numpy import linalg as LA\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from mpl_toolkits.mplot3d import Axes3D\n",
    "\n",
    "style.use(\"ggplot\")\n",
    "conn = sqlite3.connect(r\"C:\\Users\\oisin\\Code\\CS185CProject\\FPA_FOD_20170508.sqlite\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_sql_query(''' SELECT * FROM classifed_weather_data ''', conn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "cdf = df.drop(['level_0', 'index', 'stn', 'wban', 'date', 'sndp'] + list(df.filter(regex='count_')) + list(df.filter(regex='flag')), axis=1).replace([9999.9, 999.9], 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "cdf = cdf.apply(pd.to_numeric, errors='coerce')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "fdf = cdf[cdf['isFire'] == 1]\n",
    "ndf = cdf[cdf['isFire'] == 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "b = fdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "mean = np.mean(b, axis=0)\n",
    "At = b - mean\n",
    "A = At.transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y = 1/b.shape[0] * A"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "U, S, V = LA.svd(Y, full_matrices=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(18, 18)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "U.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((18,), array([1.19280107e+01, 7.93783127e+00, 1.53468067e+00, 5.95217563e-01,\n",
       "        4.37528045e-01, 2.77598099e-01, 2.39838401e-01, 2.13228496e-01,\n",
       "        1.56634903e-01, 8.97974190e-02, 5.02092987e-02, 4.43344832e-03,\n",
       "        3.31632288e-03, 1.77321767e-03, 1.18091849e-03, 0.00000000e+00,\n",
       "        0.00000000e+00, 0.00000000e+00]))"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "S.shape, S"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "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": [
    "ypos = np.arange(len(fdf.columns))\n",
    "plt.bar(ypos, S)\n",
    "plt.xticks(ypos, fdf.columns, rotation='90')\n",
    "plt.title(\"Eigenvalues for each feature\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['temp', 'dewp', 'slp'], dtype='object')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "important_cols = fdf.columns[np.where(S > 1)]\n",
    "important_cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_cdf = pd.concat([fdf.sample(n=2000, random_state=1), ndf.sample(n=98000, random_state=1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = test_cdf[important_cols].values\n",
    "Y = test_cdf['isFire'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = scaler.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "          intercept_scaling=1, loss='squared_hinge', max_iter=500000,\n",
       "          multi_class='ovr', penalty='l2', random_state=1, tol=0.0001,\n",
       "          verbose=0)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = LinearSVC(random_state=1, max_iter=500000)\n",
    "clf.fit(X, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xt = X.transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.06147369, -0.0482488 , -0.00775867]])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "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": [
    "fig = plt.figure()\n",
    "ax = plt.axes(projection=\"3d\")\n",
    "color = ['white' if c == 0 else 'red' for c in Y]\n",
    "ax.scatter(Xt[0], Xt[1], Xt[2], c=color, alpha=0.1)\n",
    "\n",
    "ax.set_xlabel(important_cols[0], fontsize = 10)\n",
    "ax.set_ylabel(important_cols[1], fontsize = 10)\n",
    "ax.set_zlabel(important_cols[2], fontsize = 10)\n",
    "\n",
    "C = lambda x,y: (-clf.intercept_[0] - clf.coef_[0][0] * x - clf.coef_[0][1] * y) / clf.coef_[0][2]\n",
    "A,B = np.meshgrid(np.linspace(min(Xt[0]),max(Xt[0]),20),np.linspace(min(Xt[1]),max(Xt[1]),20))\n",
    "ax.plot_surface(A, B, C(A,B), color='yellow', alpha=0.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing Accuracy of SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sklearn.metrics as metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_data = pd.concat([fdf.sample(n=1000, random_state=3), ndf.sample(n=5000, random_state=3)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test = test_data['isFire']\n",
    "x_test = test_data[important_cols].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 1, ..., 0, 1, 0], dtype=int64)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction = clf.predict(x_test)\n",
    "prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.72      0.79      5000\n",
      "           1       0.24      0.45      0.31      1000\n",
      "\n",
      "    accuracy                           0.68      6000\n",
      "   macro avg       0.55      0.58      0.55      6000\n",
      "weighted avg       0.76      0.68      0.71      6000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, prediction))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average precision-recall score: 0.20\n"
     ]
    }
   ],
   "source": [
    "average_precision = metrics.average_precision_score(y_test, prediction)\n",
    "\n",
    "print('Average precision-recall score: {0:0.2f}'.format(\n",
    "      average_precision))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '2-class Precision-Recall curve: AP=0.20')"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "from inspect import signature\n",
    "\n",
    "precision, recall, _ = metrics.precision_recall_curve(y_test, prediction)\n",
    "\n",
    "# In matplotlib < 1.5, plt.fill_between does not have a 'step' argument\n",
    "step_kwargs = ({'step': 'post'}\n",
    "               if 'step' in signature(plt.fill_between).parameters\n",
    "               else {})\n",
    "plt.step(recall, precision, color='b', alpha=0.2,\n",
    "         where='post')\n",
    "plt.fill_between(recall, precision, alpha=0.2, color='b', **step_kwargs)\n",
    "\n",
    "plt.xlabel('Recall')\n",
    "plt.ylabel('Precision')\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.title('2-class Precision-Recall curve: AP={0:0.2f}'.format(\n",
    "          average_precision))"
   ]
  }
 ],
 "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": "3.7.3"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": true,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}