Logistic Regression.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "from time import time\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from scipy.stats import randint as sp_randint\n",
    "from seaborn import pairplot\n",
    "\n",
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "from sklearn.linear_model import ElasticNet\n",
    "from sklearn.linear_model import ElasticNet, ElasticNetCV\n",
    "from sklearn.linear_model import Lasso\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.linear_model import Ridge\n",
    "\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.metrics import r2_score\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from sklearn.utils import shuffle\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.svm import LinearSVC\n",
    "\n",
    "import mglearn\n",
    "\n",
    "from sklearn.datasets import load_iris\n",
    "import random\n",
    "random.seed(15)\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "import seaborn as sns\n",
    "\n",
    "from collections import Counter\n",
    "\n",
    "from imblearn.over_sampling import RandomOverSampler\n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>female</th>\n",
       "      <th>flu_season</th>\n",
       "      <th>ed_admit</th>\n",
       "      <th>severity score</th>\n",
       "      <th>comorbidity score</th>\n",
       "      <th>readmit30</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4377</th>\n",
       "      <td>88</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>34</td>\n",
       "      <td>94</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4378</th>\n",
       "      <td>98</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>51</td>\n",
       "      <td>136</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4379</th>\n",
       "      <td>84</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>32</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4380</th>\n",
       "      <td>67</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>73</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4381</th>\n",
       "      <td>79</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>92</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      age  female  flu_season  ed_admit  severity score  comorbidity score  \\\n",
       "4377   88       1           0         1              34                 94   \n",
       "4378   98       0           0         1              51                136   \n",
       "4379   84       1           0         1              10                 32   \n",
       "4380   67       1           1         1              10                 73   \n",
       "4381   79       1           0         1              20                 92   \n",
       "\n",
       "      readmit30  \n",
       "4377          0  \n",
       "4378          0  \n",
       "4379          0  \n",
       "4380          0  \n",
       "4381          0  "
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# admissions over last year for patients with AMI\n",
    "\n",
    "# severity score - generic physiologic severity of illness score based on lab tests and vital signs\n",
    "# comorbidity score - severity score based on patients' pre-existing diagnoses\n",
    "# readmit30 - indicator for hospital readmission within 30 day\n",
    "\n",
    "df = pd.read_csv('Final Project Data_Case.csv').dropna()\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6768, 6)\n",
      "X_train shape: (5076, 6)\n",
      "y_train shape: (5076,)\n",
      "X_test shape: (1692, 6)\n",
      "y_test shape: (1692,)\n"
     ]
    }
   ],
   "source": [
    "X = df.ix[:,:6]\n",
    "y = df['readmit30']\n",
    "\n",
    "# up-sample: resample the minority class to deal with class imbalance\n",
    "\n",
    "# !pip install -U imbalanced-learn\n",
    "# from imblearn.over_sampling import RandomOverSampler\n",
    "\n",
    "ros = RandomOverSampler()\n",
    "X_resampled, y_resampled = ros.fit_sample(X, y)\n",
    "\n",
    "print(X_resampled.shape)\n",
    "\n",
    "# default 75-25 train test split\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X_resampled, y_resampled, random_state=0)\n",
    "\n",
    "print(\"X_train shape: {}\".format(X_train.shape))\n",
    "print(\"y_train shape: {}\".format(y_train.shape))\n",
    "print(\"X_test shape: {}\".format(X_test.shape))\n",
    "print(\"y_test shape: {}\".format(y_test.shape))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x11209ee10>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkoAAADNCAYAAABHCV/QAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x111f15470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# train classifiers using severity score comorbidity score\n",
    "\n",
    "# X = np.array(X_train[['severity score', 'comorbidity score']])\n",
    "X = np.array(X_train[:,4:6])\n",
    "y = np.array(y_train)\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(10, 3))\n",
    "for model, ax in zip([LinearSVC(), LogisticRegression()], axes):\n",
    "    clf = model.fit(X, y)\n",
    "    \n",
    "    # visualize decision boundary found by linear model\n",
    "    mglearn.plots.plot_2d_separator(clf, X, fill=False, eps=0.5,\n",
    "    ax=ax, alpha=.7)\n",
    "    mglearn.discrete_scatter(X[:, 0], X[:, 1], y, ax=ax)\n",
    "    ax.set_title(\"{}\".format(clf.__class__.__name__))\n",
    "    ax.set_xlabel(\"severity score\")\n",
    "    ax.set_ylabel(\"comorbidity score\")\n",
    "axes[0].legend()\n",
    "\n",
    "# Decision boundaries of a linear SVM and logistic regression on the dataset with the default parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set score: 0.98\n",
      "Test set score: 0.90\n"
     ]
    }
   ],
   "source": [
    "# now use all the features\n",
    "\n",
    "clf = SVC() # use default values for C and gamma\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "y_pred = clf.predict(X_test)\n",
    "\n",
    "# Returns the mean accuracy on the given test data and labels.\n",
    "\n",
    "print(\"Training set score: {:.2f}\".format(clf.score(X_train, y_train)))\n",
    "print(\"Test set score: {:.2f}\".format(clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set score: 1.00\n",
      "Test set score: 0.89\n"
     ]
    }
   ],
   "source": [
    "# now use all the features\n",
    "\n",
    "clf = RandomForestClassifier() # use default values for C and gamma\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "y_pred = clf.predict(X_test)\n",
    "\n",
    "# Returns the mean accuracy on the given test data and labels.\n",
    "\n",
    "print(\"Training set score: {:.2f}\".format(clf.score(X_train, y_train)))\n",
    "print(\"Test set score: {:.2f}\".format(clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set score: 0.71\n",
      "Test set score: 0.71\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "# now use all the features\n",
    "\n",
    "clf = LogisticRegression() # use default values for C and gamma\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "y_pred = clf.predict(X_test)\n",
    "\n",
    "# Returns the mean accuracy on the given test data and labels.\n",
    "\n",
    "print(\"Training set score: {:.2f}\".format(clf.score(X_train, y_train)))\n",
    "print(\"Test set score: {:.2f}\".format(clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}