KAGGLE - Heart Disease UCI (Dataset).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# KAGGLE - Heart Disease UCI (Dataset)\n",
    "\n",
    "Dataset obtained from https://www.kaggle.com/ronitf/heart-disease-uci\n",
    "\n",
    "Before we start working on the dataset, it is useful to try and visualise it a bit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['ï»¿age' 'sex' 'cp' 'trestbps' 'chol' 'fbs' 'restecg' 'thalach' 'exang'\n",
      " 'oldpeak' 'slope' 'ca' 'thal' 'target']\n",
      "[[63.  1.  3. ...  0.  1.  1.]\n",
      " [37.  1.  2. ...  0.  2.  1.]\n",
      " [41.  0.  1. ...  0.  2.  1.]\n",
      " ...\n",
      " [68.  1.  0. ...  2.  3.  0.]\n",
      " [57.  1.  0. ...  1.  3.  0.]\n",
      " [57.  0.  1. ...  1.  2.  0.]]\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.patches import Ellipse\n",
    "\n",
    "#importing all of the data as a numpy array since it's a small dataset (~12KB)\n",
    "data_set_labels = np.genfromtxt(\"heart-disease-uci/heart.csv\",dtype=np.str,delimiter=\",\")[0]\n",
    "data_set_values = np.genfromtxt(\"heart-disease-uci/heart.csv\",dtype=np.float,delimiter=\",\",skip_header=1)\n",
    "print(data_set_labels)\n",
    "print(data_set_values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data contains the following variables: \n",
    "\n",
    "- `age`: age in years \n",
    "- `sex`: male(1)/female(0) \n",
    "- `cp`: chest pain type (1,2,3,4) \n",
    "- `trestbps`: resting blood pressure (in mmHg as measured on admission to the hospital) \n",
    "- `chol`: serum cholestoral in mg/dl \n",
    "- `fbs`: (fasting blood sugar > 120 mg/dl) (true(1)/false(0))\n",
    "- `restecg`: resting electrocardiographic results \n",
    "- `thalach`: maximum heart rate achieved \n",
    "- `exang`: exercise induced angina (yes(1)/no(0))\n",
    "- `oldpeak`: ST depression induced by exercise relative to rest \n",
    "- `slope`: the slope of the peak exercise ST segment \n",
    "- `ca`: number of major vessels (0,1,2,3) colored by flourosopy \n",
    "- `thal`: normal(3)/fixed defect(6)/reversable defect(7) \n",
    "- `target`: have disease or not (yes(1)/no(0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 207 males and 96 females\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ca51d86668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 138 healthy and 165 sick\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ca53e26eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 258 with low and 45 high fasting blood sugar\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAADTxJREFUeJzt3X+s3fVdx/HnaxQxETLA3hIsxYukujGTdcsNEkkUxDDolhRiMJC51UnSLStxxM3YLSZDDdoZYckSh3YZWZewIWYsNFKnWDE4lB+3gPzqCM1a4dKmvdt0MKdMurd/3G/dodzee+4999zbfvp8JDfnnM/5fM953+TmeQ/fnnNJVSFJateblnoASdJwGXpJapyhl6TGGXpJapyhl6TGGXpJapyhl6TGGXpJapyhl6TGLVvqAQCWL19eo6OjSz2GJB1Xdu7c+a2qGplt3zER+tHRUcbHx5d6DEk6riT59372eepGkhpn6CWpcYZekhpn6CWpcYZekhpn6CWpcYZekhpn6CWpcYZekhp3THwydhCjm+5d6hF0DNu7+d1LPYK05HxFL0mNmzX0SVYluT/JriTPJPlIt35TkpeSPNF9re055uNJdid5Lsm7hvkNSJJm1s+pm9eAj1bVY0lOA3Ymua+779NV9We9m5NcAFwLvA34KeAfkvxsVR1ayMElSf2Z9RV9Ve2vqse6668Au4CVMxyyDrizql6tqj3AbuDChRhWkjR3czpHn2QUeAfwcLd0Q5Ink9ye5IxubSXwYs9hE8z8i0GSNER9hz7JqcBXgBur6mXgNuB8YA2wH7jl8NZpDq9pHm9DkvEk45OTk3MeXJLUn75Cn+RkpiJ/R1XdDVBVB6rqUFX9EPgcPzo9MwGs6jn8HGDfkY9ZVVuqaqyqxkZGZv0fpEiS5qmfd90E+Dywq6pu7Vk/u2fb1cDT3fVtwLVJTklyHrAaeGThRpYkzUU/77q5GHgf8FSSJ7q1TwDXJVnD1GmZvcAHAarqmSR3Ac8y9Y6djb7jRpKWzqyhr6qvM/159+0zHHMzcPMAc0mSFoifjJWkxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWrcrKFPsirJ/Ul2JXkmyUe69TOT3Jfk+e7yjG49ST6TZHeSJ5O8c9jfhCTp6Pp5Rf8a8NGqeitwEbAxyQXAJmBHVa0GdnS3Aa4EVndfG4DbFnxqSVLfZg19Ve2vqse6668Au4CVwDpga7dtK3BVd30d8MWa8hBwepKzF3xySVJf5nSOPsko8A7gYeCsqtoPU78MgBXdtpXAiz2HTXRrRz7WhiTjScYnJyfnPrkkqS99hz7JqcBXgBur6uWZtk6zVm9YqNpSVWNVNTYyMtLvGJKkOeor9ElOZiryd1TV3d3ygcOnZLrLg936BLCq5/BzgH0LM64kaa76eddNgM8Du6rq1p67tgHru+vrgXt61t/fvfvmIuC7h0/xSJIW37I+9lwMvA94KskT3dongM3AXUmuB14Arunu2w6sBXYD3wc+sKATS5LmZNbQV9XXmf68O8Bl0+wvYOOAc0mSFoifjJWkxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxs0a+iS3JzmY5OmetZuSvJTkie5rbc99H0+yO8lzSd41rMElSf3p5xX9F4Arpln/dFWt6b62AyS5ALgWeFt3zGeTnLRQw0qS5m7W0FfVA8B3+ny8dcCdVfVqVe0BdgMXDjCfJGlAg5yjvyHJk92pnTO6tZXAiz17Jrq1N0iyIcl4kvHJyckBxpAkzWS+ob8NOB9YA+wHbunWM83emu4BqmpLVY1V1djIyMg8x5AkzWZeoa+qA1V1qKp+CHyOH52emQBW9Ww9B9g32IiSpEHMK/RJzu65eTVw+B0524Brk5yS5DxgNfDIYCNKkgaxbLYNSb4MXAIsTzIBfBK4JMkapk7L7AU+CFBVzyS5C3gWeA3YWFWHhjO6JKkfs4a+qq6bZvnzM+y/Gbh5kKEkSQvHT8ZKUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuNmDX2S25McTPJ0z9qZSe5L8nx3eUa3niSfSbI7yZNJ3jnM4SVJs+vnFf0XgCuOWNsE7Kiq1cCO7jbAlcDq7msDcNvCjClJmq9ZQ19VDwDfOWJ5HbC1u74VuKpn/Ys15SHg9CRnL9SwkqS5m+85+rOqaj9Ad7miW18JvNizb6JbkyQtkYX+x9hMs1bTbkw2JBlPMj45ObnAY0iSDptv6A8cPiXTXR7s1ieAVT37zgH2TfcAVbWlqsaqamxkZGSeY0iSZjPf0G8D1nfX1wP39Ky/v3v3zUXAdw+f4pEkLY1ls21I8mXgEmB5kgngk8Bm4K4k1wMvANd027cDa4HdwPeBDwxhZknSHMwa+qq67ih3XTbN3gI2DjqUJGnh+MlYSWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWqcoZekxhl6SWrcskEOTrIXeAU4BLxWVWNJzgT+ChgF9gK/XlX/MdiYkqT5WohX9JdW1ZqqGutubwJ2VNVqYEd3W5K0RIZx6mYdsLW7vhW4agjPIUnq06ChL+Dvk+xMsqFbO6uq9gN0lysGfA5J0gAGOkcPXFxV+5KsAO5L8o1+D+x+MWwAOPfccwccQ5J0NAO9oq+qfd3lQeCrwIXAgSRnA3SXB49y7JaqGquqsZGRkUHGkCTNYN6hT/ITSU47fB24HHga2Aas77atB+4ZdEhJ0vwNcurmLOCrSQ4/zpeq6mtJHgXuSnI98AJwzeBjSpLma96hr6pvAm+fZv3bwGWDDCW1ZnTTvUs9go5Reze/e+jP4SdjJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxhl6SGmfoJalxQwt9kiuSPJdkd5JNw3oeSdLMhhL6JCcBfw5cCVwAXJfkgmE8lyRpZsN6RX8hsLuqvllVPwDuBNYN6bkkSTMYVuhXAi/23J7o1iRJi2zZkB4306zV6zYkG4AN3c3vJXluSLOcaJYD31rqIY4V+dRST6Bp+DPaY8Cf0Z/uZ9OwQj8BrOq5fQ6wr3dDVW0Btgzp+U9YScaramyp55COxp/RxTesUzePAquTnJfkx4BrgW1Dei5J0gyG8oq+ql5LcgPwd8BJwO1V9cwwnkuSNLNhnbqhqrYD24f1+DoqT4fpWOfP6CJLVc2+S5J03PJPIEhS4wz9cSDJbyfZleSOo9x/SZK/Wey5pF5JTk/y4aWeQ29k6I8PHwbWVtV7l3oQaQanM/Wz+jrdn0TREjL0x7gkfwH8DLAtye8l+Zckj3eXPzfN/l9O8kT39XiS07r1303yaJInk/zBYn8fOiFsBs7vfvYeTXJ/ki8BTyUZTfL04Y1JPpbkpu76+Um+lmRnkn9O8pYlmr9ZQ3vXjRZGVX0oyRXApcAPgFu6t6/+KvDHwK8dccjHgI1V9WCSU4H/SXI5sJqpv0EUpn5p/FJVPbB434lOAJuAn6+qNUkuAe7tbu9JMjrDcVuAD1XV80l+Afgs8CvDHvZEYuiPL28GtiZZzdSflDh5mj0PArd25/PvrqqJLvSXA493e05lKvyGXsP0SFXtmWlD92LkF4G/Tv7/L6ecMuzBTjSG/vjyR8D9VXV19wrpn47cUFWbk9wLrAUe6l75B/iTqvrLRZxV+q+e66/x+lPFP95dvgn4z6pas2hTnYA8R398eTPwUnf9N6fbkOT8qnqqqj4FjANvYeoTyr/VvXoiycokKxZhXp1YXgFOO8p9B4AVSX4yySnAewCq6mVgT5JrADLl7Ysy7QnEV/THlz9l6tTN7wD/eJQ9Nya5FDgEPAv8bVW9muStwL92/3n8PeA3gIOLMLNOEFX17SQPdv/o+t9Mxf3wff+b5A+Bh4E9wDd6Dn0vcFuS32fqdOSdwL8t3uTt85OxktQ4T91IUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ1ztBLUuMMvSQ17v8AQQntX3+gyeUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ca53e219e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Scatter plot showing the heart rate/blood pressure values for healthy and sick people\n",
      "(Ellipses show 1-sigma of the distribution)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ca51d11860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ca53e9ba90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ca53e77668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Bar chart showing male-female split\n",
    "male_female_split = np.array([np.sum(data_set_values[:,1]),data_set_values.shape[0]-np.sum(data_set_values[:,1])])\n",
    "print(\"There are {} males and {} females\".format(int(male_female_split[0]),int(male_female_split[1])))\n",
    "plt.bar([\"male\",\"female\"],male_female_split)\n",
    "plt.show()\n",
    "\n",
    "#Bar chart showing sick-healthy split\n",
    "sick_healthy_split = np.array([data_set_values.shape[0]-np.sum(data_set_values[:,-1]),np.sum(data_set_values[:,-1])])\n",
    "print(\"There are {} healthy and {} sick\".format(int(sick_healthy_split[0]),int(sick_healthy_split[1])))\n",
    "plt.bar([\"healthy\",\"sick\"],male_female_split)\n",
    "plt.show()\n",
    "\n",
    "#Bar chart showing low-high fasting blood sugar split\n",
    "fbs_split = np.array([data_set_values.shape[0]-np.sum(data_set_values[:,5]),np.sum(data_set_values[:,5])])\n",
    "print(\"There are {} with low and {} high fasting blood sugar\".format(int(fbs_split[0]),int(fbs_split[1])))\n",
    "plt.bar([\"false\",\"true\"],fbs_split)\n",
    "plt.show()\n",
    "\n",
    "#Scatter plot showing the heart rate/blood pressure values for healthy and sick people\n",
    "print(\"Scatter plot showing the heart rate/blood pressure values for healthy and sick people\\n(Ellipses show 1-sigma of the distribution)\")\n",
    "\n",
    "thalach_trestbps = np.array([data_set_values[:,3],data_set_values[:,7],data_set_values[:,-1]]).T\n",
    "healthy_thalach_trestbps = np.array([])\n",
    "sick_thalach_trestbps = np.array([])\n",
    "for i in range(data_set_values.shape[0]):\n",
    "    if thalach_trestbps[i,-1]==0:\n",
    "        healthy_thalach_trestbps = np.append(healthy_thalach_trestbps,thalach_trestbps[i,:])\n",
    "    else:\n",
    "        sick_thalach_trestbps = np.append(sick_thalach_trestbps,thalach_trestbps[i,:])\n",
    "\n",
    "healthy_thalach_trestbps = healthy_thalach_trestbps.reshape((int(sick_healthy_split[0]),3))[:,0:2]\n",
    "sick_thalach_trestbps = sick_thalach_trestbps.reshape((int(sick_healthy_split[1]),3))[:,0:2]\n",
    "\n",
    "avg_healthy_thalach_trestbps = np.average(healthy_thalach_trestbps,axis=0)\n",
    "avg_sick_thalach_trestbps = np.average(sick_thalach_trestbps,axis=0)\n",
    "std_healthy_thalach_trestbps = np.std(healthy_thalach_trestbps,axis=0)\n",
    "std_sick_thalach_trestbps = np.std(sick_thalach_trestbps,axis=0)\n",
    "\n",
    "ax = plt.gca()\n",
    "ax.add_patch(Ellipse(xy=(avg_healthy_thalach_trestbps[0],avg_healthy_thalach_trestbps[1]), width=std_healthy_thalach_trestbps[0], height=std_healthy_thalach_trestbps[1],alpha=0.45,facecolor=\"green\"))\n",
    "ax.add_patch(Ellipse(xy=(avg_sick_thalach_trestbps[0],avg_sick_thalach_trestbps[1]), width=std_sick_thalach_trestbps[0], height=std_sick_thalach_trestbps[1],alpha=0.45,facecolor=\"red\"))\n",
    "plt.scatter(healthy_thalach_trestbps[:,0],healthy_thalach_trestbps[:,1],c=\"green\",zorder=10)\n",
    "plt.scatter(sick_thalach_trestbps[:,0],sick_thalach_trestbps[:,1],c=\"red\",zorder=10)\n",
    "plt.ylabel(\"maximum heart rate achieved\")\n",
    "plt.xlabel(\"resting blood pressure\")\n",
    "plt.title(\"sick/healthy\")\n",
    "plt.show()\n",
    "\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "\n",
    "age_thalach = np.array([data_set_values[:,0],data_set_values[:,7],data_set_values[:,-1]]).T\n",
    "\n",
    "line = LinearRegression()\n",
    "line.fit(age_thalach[:,0].reshape((len(age_thalach[:,0]),1)), age_thalach[:,1])\n",
    "\n",
    "healthy_age_thalach = np.array([])\n",
    "sick_age_thalach = np.array([])\n",
    "for i in range(data_set_values.shape[0]):\n",
    "    if age_thalach[i,-1]==0:\n",
    "        healthy_age_thalach = np.append(healthy_age_thalach,age_thalach[i,:])\n",
    "    else:\n",
    "        sick_age_thalach = np.append(sick_age_thalach,age_thalach[i,:])\n",
    "\n",
    "healthy_age_thalach = healthy_age_thalach.reshape((int(sick_healthy_split[0]),3))[:,0:2]\n",
    "sick_age_thalach = sick_age_thalach.reshape((int(sick_healthy_split[1]),3))[:,0:2]\n",
    "\n",
    "lineH = LinearRegression()\n",
    "lineH.fit(healthy_age_thalach[:,0].reshape((len(healthy_age_thalach[:,0]),1)), healthy_age_thalach[:,1])\n",
    "\n",
    "lineS = LinearRegression()\n",
    "lineS.fit(sick_age_thalach[:,0].reshape((len(sick_age_thalach[:,0]),1)), sick_age_thalach[:,1])\n",
    "\n",
    "plt.plot(np.array([[30],[75]]),line.predict(np.array([[30],[75]])),c=\"black\")\n",
    "plt.plot(np.array([[30],[75]]),lineH.predict(np.array([[30],[75]])),c=\"green\")\n",
    "plt.plot(np.array([[30],[75]]),lineS.predict(np.array([[30],[75]])),c=\"red\")\n",
    "plt.scatter(healthy_age_thalach[:,0],healthy_age_thalach[:,1],c=\"green\",zorder=10)\n",
    "plt.scatter(sick_age_thalach[:,0],sick_age_thalach[:,1],c=\"red\",zorder=10)\n",
    "plt.ylabel(\"maximum heart rate achieved\")\n",
    "plt.xlabel(\"age\")\n",
    "plt.title(\"All: R^2 = {}\\nHealthy: R^2 = {}\\nSick: R^2 = {}\".format(r2_score(age_thalach[:,1],line.predict(age_thalach[:,0].reshape((len(age_thalach[:,0]),1)))),r2_score(healthy_age_thalach[:,1],line.predict(healthy_age_thalach[:,0].reshape((len(healthy_age_thalach[:,0]),1)))),r2_score(sick_age_thalach[:,1],line.predict(sick_age_thalach[:,0].reshape((len(sick_age_thalach[:,0]),1))))))\n",
    "plt.show()\n",
    "\n",
    "age_thalach_fbs = np.array([data_set_values[:,0],data_set_values[:,7],data_set_values[:,5]]).T\n",
    "\n",
    "lo = np.array([])\n",
    "hi = np.array([])\n",
    "for i in range(data_set_values.shape[0]):\n",
    "    if age_thalach_fbs[i,-1]==0:\n",
    "        lo = np.append(lo,age_thalach_fbs[i,:])\n",
    "    else:\n",
    "        hi = np.append(hi,age_thalach_fbs[i,:])\n",
    "\n",
    "lo = lo.reshape((int(fbs_split[0]),3))[:,0:2]\n",
    "hi = hi.reshape((int(fbs_split[1]),3))[:,0:2]\n",
    "\n",
    "plt.scatter(lo[:,0],lo[:,1],c=\"green\",zorder=10)\n",
    "plt.scatter(hi[:,0],hi[:,1],c=\"red\",zorder=10)\n",
    "plt.ylabel(\"maximum heart rate achieved\")\n",
    "plt.xlabel(\"age\")\n",
    "plt.title(\"low/high blood sugar\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The blood sugar scatter plot was plotted to check if the scientifically supported correlation of maximum heart rate and age (the Harris–Benedict equation relates age with BMR) with caloric burn.\n",
    "\n",
    "As a starting off point, we will try a SVM classifying using the heart rate/blood pressure values, heart rate/blood pressure/age/cholesterol values, and all values (the first two options are there to test if the input space can be minimised). The choice for SVMs was made since they tend to perform very well on relatively small datasets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Testing SVM accuracy with heart rate/blood pressure values\n",
      "\n",
      "Accuracy (linear): 0.7362637362637363\n",
      "Accuracy (rbf): 0.6043956043956044\n",
      "Accuracy (sigmoid): 0.5604395604395604\n",
      "\n",
      "Testing SVM accuracy with heart rate/blood pressure/age/cholesterol values\n",
      "\n",
      "Accuracy (linear): 0.7377049180327869\n",
      "Accuracy (rbf): 0.5901639344262295\n",
      "Accuracy (sigmoid): 0.5901639344262295\n",
      "\n",
      "Testing SVM accuracy with all values\n",
      "\n",
      "Accuracy (linear): 0.9010989010989011\n",
      "Accuracy (rbf): 0.5714285714285714\n",
      "Accuracy (sigmoid): 0.5604395604395604\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import svm\n",
    "from sklearn import metrics\n",
    "\n",
    "print(\"Testing SVM accuracy with heart rate/blood pressure values\\n\")\n",
    "# Split dataset into training set and test set\n",
    "X_train, X_test, y_train, y_test = train_test_split(thalach_trestbps[:,1:-1], thalach_trestbps[:,-1], test_size=0.3,random_state=109) # 70% training and 30% test\n",
    "\n",
    "#Create a linear svm Classifier\n",
    "clf = svm.SVC(kernel='linear')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (linear):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "#Create a rbf svm Classifier\n",
    "clf = svm.SVC(kernel='rbf')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (rbf):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "#Create a sigmoid svm Classifier\n",
    "clf = svm.SVC(kernel='sigmoid')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (sigmoid):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "print(\"\\nTesting SVM accuracy with heart rate/blood pressure/age/cholesterol values\\n\")\n",
    "\n",
    "#attempt with more variables\n",
    "age_trestbps_chol_thalach = np.array([data_set_values[:,0],data_set_values[:,3],data_set_values[:,4],data_set_values[:,7],data_set_values[:,-1]]).T\n",
    "\n",
    "# Split dataset into training set and test set\n",
    "X_train, X_test, y_train, y_test = train_test_split(age_trestbps_chol_thalach[:,1:-1], age_trestbps_chol_thalach[:,-1], test_size=0.2,random_state=109) # 70% training and 30% test\n",
    "\n",
    "#Create a linear svm Classifier\n",
    "clf = svm.SVC(kernel='linear')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (linear):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "#Create a rbf svm Classifier\n",
    "clf = svm.SVC(kernel='rbf')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (rbf):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "#Create a sigmoid svm Classifier\n",
    "clf = svm.SVC(kernel='sigmoid')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (sigmoid):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "print(\"\\nTesting SVM accuracy with all values\\n\")\n",
    "\n",
    "#attempt with all variables\n",
    "\n",
    "# Split dataset into training set and test set\n",
    "X_train, X_test, y_train, y_test = train_test_split(data_set_values[:,1:-1], data_set_values[:,-1], test_size=0.3,random_state=109) # 70% training and 30% test\n",
    "\n",
    "#Create a linear svm Classifier\n",
    "clf = svm.SVC(kernel='linear')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (linear):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "#Create a rbf svm Classifier\n",
    "clf = svm.SVC(kernel='rbf')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (rbf):\",metrics.accuracy_score(y_test, y_pred))\n",
    "\n",
    "#Create a sigmoid svm Classifier\n",
    "clf = svm.SVC(kernel='sigmoid')\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (sigmoid):\",metrics.accuracy_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Clearly, the Linear-SVM with all values greatly superseeded the rest with an accuracy of ~90.1%. For comparison, we will try a Neural Network with layer sizes 100,100,100 and we should expect to get a rather low accuracy given the small amount of data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Testing MLP with all values\n",
      "Accuracy (MLP): 0.7213114754098361\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neural_network import MLPClassifier\n",
    "\n",
    "print(\"Testing MLP with all values\")\n",
    "# Split dataset into training set and test set\n",
    "X_train, X_test, y_train, y_test = train_test_split(data_set_values[:,1:-1], data_set_values[:,-1], test_size=0.2,random_state=109) # 70% training and 30% test\n",
    "\n",
    "clf = MLPClassifier(max_iter=2000, activation='logistic',solver='lbfgs', alpha=1e-5,hidden_layer_sizes=(100, 100, 100), random_state=1)\n",
    "clf.fit(X_train, y_train)\n",
    "y_pred = clf.predict(X_test)\n",
    "print(\"Accuracy (MLP):\",metrics.accuracy_score(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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}