M2 - knn and imputing missing values.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## KNN and imputing missing values\n",
    "- https://harvard-iacs.github.io/2018-CS109A/lectures/lecture-13/demo/\n",
    "- https://github.com/cs109/a-2017/blob/master/Labs/Lab7_Logistic_ROC_Imputation/lab7_Logistics_ROC_Imputation_Solutions.ipynb\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pylab as plt\n",
    "import seaborn as sns\n",
    "import statsmodels.api as sm\n",
    "import sklearn.linear_model as lm\n",
    "import sklearn as sk\n",
    "import sklearn.neighbors as neighbors\n",
    "\n",
    "sns.set()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>Age</th>\n",
       "      <th>Sex</th>\n",
       "      <th>ChestPain</th>\n",
       "      <th>RestBP</th>\n",
       "      <th>Chol</th>\n",
       "      <th>Fbs</th>\n",
       "      <th>RestECG</th>\n",
       "      <th>MaxHR</th>\n",
       "      <th>ExAng</th>\n",
       "      <th>Oldpeak</th>\n",
       "      <th>Slope</th>\n",
       "      <th>Ca</th>\n",
       "      <th>Thal</th>\n",
       "      <th>AHD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>63</td>\n",
       "      <td>1</td>\n",
       "      <td>typical</td>\n",
       "      <td>145</td>\n",
       "      <td>233</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>150</td>\n",
       "      <td>0</td>\n",
       "      <td>2.3</td>\n",
       "      <td>3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>fixed</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>67</td>\n",
       "      <td>1</td>\n",
       "      <td>asymptomatic</td>\n",
       "      <td>160</td>\n",
       "      <td>286</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>108</td>\n",
       "      <td>1</td>\n",
       "      <td>1.5</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "      <td>normal</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>67</td>\n",
       "      <td>1</td>\n",
       "      <td>asymptomatic</td>\n",
       "      <td>120</td>\n",
       "      <td>229</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>129</td>\n",
       "      <td>1</td>\n",
       "      <td>2.6</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>reversable</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>nonanginal</td>\n",
       "      <td>130</td>\n",
       "      <td>250</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>187</td>\n",
       "      <td>0</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>normal</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>41</td>\n",
       "      <td>0</td>\n",
       "      <td>nontypical</td>\n",
       "      <td>130</td>\n",
       "      <td>204</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>172</td>\n",
       "      <td>0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>normal</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  Age  Sex     ChestPain  RestBP  Chol  Fbs  RestECG  MaxHR  \\\n",
       "0           1   63    1       typical     145   233    1        2    150   \n",
       "1           2   67    1  asymptomatic     160   286    0        2    108   \n",
       "2           3   67    1  asymptomatic     120   229    0        2    129   \n",
       "3           4   37    1    nonanginal     130   250    0        0    187   \n",
       "4           5   41    0    nontypical     130   204    0        2    172   \n",
       "\n",
       "   ExAng  Oldpeak  Slope   Ca        Thal  AHD  \n",
       "0      0      2.3      3  0.0       fixed   No  \n",
       "1      1      1.5      2  3.0      normal  Yes  \n",
       "2      1      2.6      2  2.0  reversable  Yes  \n",
       "3      0      3.5      3  0.0      normal   No  \n",
       "4      0      1.4      1  0.0      normal   No  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_heart = pd.read_csv('./data/Heart.csv')\n",
    "df_heart.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_x = df_heart['MaxHR'].values.reshape(-1,1)\n",
    "data_y = df_heart['AHD'].map(lambda x: 0 if x == 'No' else 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Heart Disease')"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import neighbors\n",
    "import sklearn.metrics as met\n",
    "\n",
    "\n",
    "knn1 = neighbors.KNeighborsClassifier(n_neighbors=1)\n",
    "knn20 = neighbors.KNeighborsClassifier(n_neighbors=10)\n",
    "knn50 = neighbors.KNeighborsClassifier(n_neighbors=50)\n",
    "\n",
    "# fit knn model\n",
    "knn1.fit(data_x, data_y)  \n",
    "knn20.fit(data_x, data_y) \n",
    "knn50.fit(data_x, data_y) \n",
    "\n",
    "# sort the x values for better line plot and then predict\n",
    "x = np.sort(data_x, axis = 0)\n",
    "yhat1 = knn1.predict(x) \n",
    "yhat20 = knn10.predict(x) \n",
    "yhat50 = knn50.predict(x) \n",
    "\n",
    "#met.confusion_matrix(data_y, yhat1)\n",
    "#knn1.score(data_x, data_y)\n",
    "\n",
    "plt.figure(figsize=(15,8))\n",
    "plt.plot(data_x, data_y, 'o', alpha = .1, color = 'black', markersize = 10)\n",
    "#plt.plot(x, yhat1, '-', alpha = 1, color = 'blue')\n",
    "plt.plot(x, yhat20, '-', alpha = 1, color = 'green')\n",
    "plt.plot(x, yhat50, '-', alpha = 1, color = 'maroon', linewidth = 5)\n",
    "plt.xlabel(\"Max HR\")\n",
    "plt.ylabel(\"Heart Disease\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.693069306930693\n",
      "0.7161716171617162\n",
      "0.7062706270627063\n"
     ]
    }
   ],
   "source": [
    "# print accuracy scores\n",
    "\n",
    "print(knn1.score(data_x, data_y))\n",
    "print(knn10.score(data_x, data_y))\n",
    "print(knn50.score(data_x, data_y))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lab 7 - imputing missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape (1551, 9)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>health</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>educ</th>\n",
       "      <th>sexornt</th>\n",
       "      <th>partyid</th>\n",
       "      <th>race</th>\n",
       "      <th>married</th>\n",
       "      <th>income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>excellent</td>\n",
       "      <td>53</td>\n",
       "      <td>male</td>\n",
       "      <td>16</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>rep</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>excellent</td>\n",
       "      <td>26</td>\n",
       "      <td>female</td>\n",
       "      <td>16</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>rep</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>good</td>\n",
       "      <td>59</td>\n",
       "      <td>male</td>\n",
       "      <td>13</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>rep</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>excellent</td>\n",
       "      <td>74</td>\n",
       "      <td>female</td>\n",
       "      <td>17</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>other</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>poor</td>\n",
       "      <td>37</td>\n",
       "      <td>female</td>\n",
       "      <td>10</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>other</td>\n",
       "      <td>white</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      health  age     sex  educ                   sexornt partyid   race  \\\n",
       "0  excellent   53    male    16  heterosexual or straight     rep  white   \n",
       "1  excellent   26  female    16  heterosexual or straight     rep  white   \n",
       "2       good   59    male    13  heterosexual or straight     rep  white   \n",
       "3  excellent   74  female    17  heterosexual or straight   other  white   \n",
       "4       poor   37  female    10  heterosexual or straight   other  white   \n",
       "\n",
       "   married   income  \n",
       "0        1  18750.0  \n",
       "1        1  18750.0  \n",
       "2        1      NaN  \n",
       "3        1      NaN  \n",
       "4        0      NaN  "
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gssdata = pd.read_csv('./data/gssdata4.csv')\n",
    "print(\"Shape\", gssdata.shape)\n",
    "gssdata.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "588"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(gssdata['income'].isnull())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6.382978723404255\n",
      "47.5177304964539\n",
      "26.692456479690524\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>health</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>educ</th>\n",
       "      <th>sexornt</th>\n",
       "      <th>partyid</th>\n",
       "      <th>race</th>\n",
       "      <th>married</th>\n",
       "      <th>income</th>\n",
       "      <th>poorhealth</th>\n",
       "      <th>goodhealth</th>\n",
       "      <th>excellenthealth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>excellent</td>\n",
       "      <td>53</td>\n",
       "      <td>male</td>\n",
       "      <td>16</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>rep</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>excellent</td>\n",
       "      <td>26</td>\n",
       "      <td>female</td>\n",
       "      <td>16</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>rep</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>good</td>\n",
       "      <td>59</td>\n",
       "      <td>male</td>\n",
       "      <td>13</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>rep</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>excellent</td>\n",
       "      <td>74</td>\n",
       "      <td>female</td>\n",
       "      <td>17</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>other</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>poor</td>\n",
       "      <td>37</td>\n",
       "      <td>female</td>\n",
       "      <td>10</td>\n",
       "      <td>heterosexual or straight</td>\n",
       "      <td>other</td>\n",
       "      <td>white</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      health  age     sex  educ                   sexornt partyid   race  \\\n",
       "0  excellent   53    male    16  heterosexual or straight     rep  white   \n",
       "1  excellent   26  female    16  heterosexual or straight     rep  white   \n",
       "2       good   59    male    13  heterosexual or straight     rep  white   \n",
       "3  excellent   74  female    17  heterosexual or straight   other  white   \n",
       "4       poor   37  female    10  heterosexual or straight   other  white   \n",
       "\n",
       "   married   income  poorhealth  goodhealth  excellenthealth  \n",
       "0        1  18750.0           0           0                1  \n",
       "1        1  18750.0           0           0                1  \n",
       "2        1      NaN           0           1                0  \n",
       "3        1      NaN           0           0                1  \n",
       "4        0      NaN           1           0                0  "
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# create separate columns\n",
    "# More complicated way to covnert factors into separate columns\n",
    "gssdata['poorhealth'] = np.where(gssdata['health'] == 'poor', 1,0)\n",
    "gssdata['goodhealth'] = np.where(gssdata['health'] == 'good', 1,0)\n",
    "gssdata['excellenthealth'] = np.where(gssdata['health'] == 'excellent', 1,0)\n",
    "\n",
    "# % of ppl with health level\n",
    "print(np.mean(gssdata['poorhealth'])*100)\n",
    "print(np.mean(gssdata['goodhealth'])*100)\n",
    "print(np.mean(gssdata['excellenthealth'])*100)\n",
    "gssdata.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>educ</th>\n",
       "      <th>married</th>\n",
       "      <th>income</th>\n",
       "      <th>health_excellent</th>\n",
       "      <th>health_fair</th>\n",
       "      <th>health_good</th>\n",
       "      <th>health_poor</th>\n",
       "      <th>sex_female</th>\n",
       "      <th>sex_male</th>\n",
       "      <th>sexornt_bisexual</th>\n",
       "      <th>sexornt_dont know</th>\n",
       "      <th>sexornt_heterosexual or straight</th>\n",
       "      <th>sexornt_homosexual or gay</th>\n",
       "      <th>partyid_dem</th>\n",
       "      <th>partyid_other</th>\n",
       "      <th>partyid_rep</th>\n",
       "      <th>race_black</th>\n",
       "      <th>race_other</th>\n",
       "      <th>race_white</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>53</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>26</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>59</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>74</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>37</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  educ  married   income  health_excellent  health_fair  health_good  \\\n",
       "0   53    16        1  18750.0                 1            0            0   \n",
       "1   26    16        1  18750.0                 1            0            0   \n",
       "2   59    13        1      NaN                 0            0            1   \n",
       "3   74    17        1      NaN                 1            0            0   \n",
       "4   37    10        0      NaN                 0            0            0   \n",
       "\n",
       "   health_poor  sex_female  sex_male  sexornt_bisexual  sexornt_dont know  \\\n",
       "0            0           0         1                 0                  0   \n",
       "1            0           1         0                 0                  0   \n",
       "2            0           0         1                 0                  0   \n",
       "3            0           1         0                 0                  0   \n",
       "4            1           1         0                 0                  0   \n",
       "\n",
       "   sexornt_heterosexual or straight  sexornt_homosexual or gay  partyid_dem  \\\n",
       "0                                 1                          0            0   \n",
       "1                                 1                          0            0   \n",
       "2                                 1                          0            0   \n",
       "3                                 1                          0            0   \n",
       "4                                 1                          0            0   \n",
       "\n",
       "   partyid_other  partyid_rep  race_black  race_other  race_white  \n",
       "0              0            1           0           0           1  \n",
       "1              0            1           0           0           1  \n",
       "2              0            1           0           0           1  \n",
       "3              1            0           0           0           1  \n",
       "4              1            0           0           0           1  "
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Convert categorical variables into dummy variables\n",
    "gssdata = pd.read_csv('./data/gssdata4.csv')\n",
    "gssdata = pd.get_dummies(gssdata)\n",
    "gssdata.head()\n",
    "#gssdata.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Method 1 - drop all missing values\n",
    "Drop all missing values and predict poor health from ```'age','educ','sex_female','partyid_dem','partyid_rep','income'```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(963, 20)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>educ</th>\n",
       "      <th>married</th>\n",
       "      <th>income</th>\n",
       "      <th>health_excellent</th>\n",
       "      <th>health_fair</th>\n",
       "      <th>health_good</th>\n",
       "      <th>health_poor</th>\n",
       "      <th>sex_female</th>\n",
       "      <th>sex_male</th>\n",
       "      <th>sexornt_bisexual</th>\n",
       "      <th>sexornt_dont know</th>\n",
       "      <th>sexornt_heterosexual or straight</th>\n",
       "      <th>sexornt_homosexual or gay</th>\n",
       "      <th>partyid_dem</th>\n",
       "      <th>partyid_other</th>\n",
       "      <th>partyid_rep</th>\n",
       "      <th>race_black</th>\n",
       "      <th>race_other</th>\n",
       "      <th>race_white</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>53</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>26</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>30</td>\n",
       "      <td>15</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>30</td>\n",
       "      <td>14</td>\n",
       "      <td>1</td>\n",
       "      <td>16250.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>38</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    age  educ  married   income  health_excellent  health_fair  health_good  \\\n",
       "0    53    16        1  18750.0                 1            0            0   \n",
       "1    26    16        1  18750.0                 1            0            0   \n",
       "5    30    15        1  18750.0                 1            0            0   \n",
       "9    30    14        1  16250.0                 1            0            0   \n",
       "12   38    15        0  18750.0                 0            0            1   \n",
       "\n",
       "    health_poor  sex_female  sex_male  sexornt_bisexual  sexornt_dont know  \\\n",
       "0             0           0         1                 0                  0   \n",
       "1             0           1         0                 0                  0   \n",
       "5             0           1         0                 0                  0   \n",
       "9             0           1         0                 0                  0   \n",
       "12            0           1         0                 0                  0   \n",
       "\n",
       "    sexornt_heterosexual or straight  sexornt_homosexual or gay  partyid_dem  \\\n",
       "0                                  1                          0            0   \n",
       "1                                  1                          0            0   \n",
       "5                                  1                          0            0   \n",
       "9                                  1                          0            0   \n",
       "12                                 1                          0            1   \n",
       "\n",
       "    partyid_other  partyid_rep  race_black  race_other  race_white  \n",
       "0               0            1           0           0           1  \n",
       "1               0            1           0           0           1  \n",
       "5               1            0           0           1           0  \n",
       "9               1            0           1           0           0  \n",
       "12              0            0           0           0           1  "
      ]
     },
     "execution_count": 215,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Two ways to drop all rows with missing values\n",
    "gssdata = gssdata.dropna(axis = 'rows')\n",
    "#gssdata = gssdata.dropna(how = 'any')\n",
    "print(gssdata.shape)\n",
    "\n",
    "gssdata.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "health_excellent    30.529595\n",
       "health_fair         16.303219\n",
       "health_good         51.505711\n",
       "health_poor          1.661475\n",
       "dtype: float64"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "100*np.mean(gssdata[['health_excellent', 'health_fair', 'health_good', 'health_poor']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score =  0.9813278008298755\n",
      "coeffients betas \n",
      " [[ 4.57902142e-03 -2.44763640e-01 -4.76004271e-01  1.17425904e+00\n",
      "   8.35128752e-03 -7.11763712e-05]]\n"
     ]
    }
   ],
   "source": [
    "# Specify the X and Y vectors\n",
    "X = gssdata[['age','educ','sex_female','partyid_dem','partyid_rep','income']]\n",
    "Y = gssdata['health_poor']\n",
    "\n",
    "# split into train and test\n",
    "from sklearn.model_selection import train_test_split\n",
    "Xtrain, Xtest, ytrain, ytest = train_test_split(X, Y, random_state=1, train_size = 0.5)\n",
    "\n",
    "# linear models from sklearn already imported as lm\n",
    "\n",
    "logit = lm.LogisticRegression()\n",
    "logit.fit(Xtrain, ytrain)\n",
    "print(\"score = \", logit.score(Xtest, ytest))\n",
    "print(\"coeffients betas \\n\", logit.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[473   0]\n",
      " [  9   0]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "\n",
    "# confusion_matrix(actual, pred)\n",
    "# Columns are for predicted (0,1) and rows are observed (0,1)\n",
    "ypred = logit.predict(Xtest)\n",
    "print(confusion_matrix(ytest, ypred))\n",
    "\n",
    "#plot_confusion_matrix(logit, Xtest, ytest)  \n",
    "\n",
    "# plot probability of predicting 1\n",
    "yhat = logit.predict_proba(Xtrain);\n",
    "plt.hist(yhat[:,1]);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[458,  15],\n",
       "       [  8,   1]], dtype=int64)"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# If you change the probability of predicting 1 from default 0.5 to .06, notice how the number of predicted 1s goes up\n",
    "confusion_matrix(ytest, yhat[:,1] > 0.06)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'True Positive Rate (TPR)')"
      ]
     },
     "execution_count": 244,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEXCAYAAACtTzM+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAABEI0lEQVR4nO3dd3hUZfbA8e+kJyQkEJLQkR5aQidAKIlAEBJK6EQBQRRRWVlFkCIrobiRFV1clboqUkSquPuDUBI6QemEJs0AQiqkt8m8vz+QWSIMk0Amk3I+z+PzcOeW97x34j3z3nKuRimlEEIIUe5ZmDsAIYQQJYMkBCGEEIAkBCGEEH+QhCCEEAKQhCCEEOIPkhCEEEIAkhDMpnHjxgQFBdG/f38GDBhAQEAAgwYN4syZM/plMjIy+Pvf/05AQABBQUEEBQWxaNEisrKy8m1r8+bNDBs2jP79+9OnTx9mzZpFSkpKcXepwGbPno2/vz+LFi0ydygFMm3aNFasWGGS7X3++efs2rXLJO08bPz48Vy+fLlItrV48WLmzJlTJNsCmDFjBocOHXriMjNnzuTs2bMFXv5hixcvxsfHh/79+9O/f3/69euHv78/CxYsQO66z8/K3AGUZ9988w2VK1fWT69YsYK5c+fy/fffo9Vqefnll2nZsiVbtmzB3t6ezMxM/vGPfzBu3Di++eYbrKys+Oqrr9i3bx//+te/qFKlCrm5ucyfP58JEyawZs0aM/bOsO+//57IyEiqVq1q7lDMLioqigYNGpi8nWXLlpm8jac1b948o8scOnSIYcOGFXj5P+vTpw8ffPCBfjo5OZl+/frh6+tLly5dCr29skoSQgmh1Wq5ffs2zs7OAGzfvh2dTsf777+vX8be3p4ZM2YwYMAAdu7cSbdu3ViyZAmbN2+mSpUqAFhbW/Pee++xc+dOcnJysLGxyddOREQEn376KTqdDgcHBz788EMcHR0JCgrixIkTANy8eVM/vWnTJjZs2EBmZiaOjo7k5uby8ssvExAQAMDHH38MwJQpU/jhhx9Yu3YtOp0OFxcXZs2aRf369fO1P3LkSJRSjB8/ntmzZ+Ps7MycOXO4d+8eGo2GsWPHMmDAAKKiopg3bx4ODg6kp6ezcePGfH3x9/enb9++HDx4kNTUVF5++WVGjhwJ3E84q1atwsLCgipVqjBr1izq1q1LamoqH374IRcuXECj0dClSxf++te/YmVlRfPmzXn++ee5cOECCxcupEWLFvniPnHiBMOHDychIYGGDRvyj3/8AwcHB65cucK8efO4d+8eeXl5vPTSSwwePBidTsf8+fM5deoU6enpKKWYO3cubdq00W9z9erVnD17lrCwMCwtLZ/YzsOmTZtGw4YNGTdu3CPTa9asYd26dVhbW2Nra8ucOXNo0KAB/v7+fPbZZ2RkZLBo0SJq1arFr7/+ilar5cMPP6RNmzYkJSXx/vvvExMTg4uLC25ubjRs2JC33nqrwH/Hv/zyC2FhYWRmZmJtbc3bb79N165dycvLIywsjD179uDk5ISXlxdXrlxh1apVvPTSS4SEhNCjRw9CQ0M5fvw41tbW1KxZkwULFrB06VLi4uJ49913CQsLY+HChYSEhNC7d+/H/j17enoajTMhIYGsrCz9/2+GvkeApUuXsmHDBipUqEDbtm3ZvXs3e/bsKfA+KVWUMItGjRqpwMBAFRgYqDp37qz8/f1VaGioSkhIUEopNWfOHPXRRx89dt0FCxao0NBQdebMGeXj41PgNuPj41WbNm1UdHS0UkqpHTt2qHHjxqkbN26oli1b6pd7eHrjxo2qXbt2KjU1VSml1IYNG9Srr76qlFJKq9UqX19fde3aNRUVFaVGjhypMjIylFJK7d+/X/Xu3dtg3xMTE1Vubq56/vnn1Y4dO5RSSt25c0d16dJFHT9+XB05ckR5enqqmzdvPnYbfn5+atasWUqn06nbt2+rDh06qAsXLqhDhw6pHj16qMTERH38L7zwgtLpdOq9995ToaGhSqfTqezsbDV27Fi1ZMkSfUybN29+bFtTp05VgwcPVhkZGUqr1aqBAweqzZs3q9zcXNWnTx919uxZpZRSKSkp6oUXXlAnTpxQx48fV2+99ZbKy8tTSim1ZMkS9dprr+m3t3z5cqWUUi+++KL6v//7vye287h4Hqz/8LRWq1XNmjVTsbGxSimlNm/erNatW6ffX6dPn1ZHjhxRTZo0UefOnVNKKbVixQoVEhKilFJq8uTJKiwsTCmlVGxsrOrcubP65z//+Uj7//znP9WHH374yOdJSUmqY8eO6uTJk0oppS5duqTat2+vYmJi1Nq1a1VISIjKysrS7/sXX3wx3z74+eefVe/evZVOp1NKKRUWFqaOHTuWL/6Hlzf09/y4eDt06KD69eunevbsqdq3b6/GjBmj3+9P+h737dunAgICVHJystLpdOr9999Xfn5+j7RRVsgIwYwenDKKjo7m1VdfpUOHDri6uurna7Xax66Xk5ODpaUlFhYW6HS6Ard3/PhxGjZsSNOmTQHo1asXvXr14ubNm09cr3Hjxjg6OgL3h95hYWHEx8dz7tw5nnvuOZ577jnWr1/Pb7/9xvDhw/XrpaSkcO/ePVxcXB673evXr5OdnU2vXr0A8PDwoFevXuzfv58OHTpQrVo1atSoYTCukSNHotFoqFq1Kl26dOHgwYMkJCTQp08f/am44OBg5s2bx82bN9m3bx9r165Fo9FgY2PD8OHD+eabb3j11VcBaNu2rcG2evTogb29PQANGzYkKSmJ69evExMTw/Tp0/XLZWVlce7cOUaOHImzszPr1q3jxo0bREVFUaFChSfuZ0PtFJSlpSW9e/dm+PDhdO/eHV9fX7p16/bIctWrV6dJkyYANG3alM2bNwOwd+9e/b/d3d3p3bt3gdsGOH36NLVr18bb21sff+vWrTl69Ch79+6lf//+2NraAjBs2DBWrVqVb/1GjRphaWnJkCFD8PX1JSAgAC8vL4PtGfp7fpwHp4xycnIIDQ3l8uXL+Pv7Azzxe7x69Sq9e/emYsWKAISEhHDkyJFC7ZfSRC4qlwDNmjXj/fffZ9q0afqDc+vWrfnll18eOeDrdDp+/vlnWrVqRYMGDdBqtVy/fj3fMtnZ2YwfP57Y2Nh8n1taWqLRaPTTSin96RP10MW13NzcfOs9fMrC3t6egIAAfvrpJzZu3MiQIUP0cfXv35+tW7eydetWNm/ezMaNG/VD8sfJy8vLF8+DmB4kwj+fKvkzK6v//Z7R6XQGE+SDbep0unzt6XS6fEn3Se093NaD/ZWXl4eTk5O+z1u3bmX9+vUMGjSIyMhIXnvtNQCef/55RowY8cS+PKmdP3vS97Vw4UK++uorateuzdKlS/nrX//6yPp2dnaP3ZaVlVW+7VpYFO7w8KTv8+F+Gdp2xYoV2bp1K1OnTsXS0pK3336b1atXG2zP0N/zk9jY2DBr1izS0tIICwvTx23oe/zzPnlwaq+skoRQQgQGBuLl5cWCBQsACAgIwN7envnz5+vvKsrKyiI0NJQKFSrQs2dPbGxsGD9+PDNmzCAhIQG4P3qYP38+mZmZeHh45GvD29ubK1eu8OuvvwKwe/dupkyZQsWKFcnNzdXfhfKf//znibEOHTqUzZs3c/z4cf21BF9fX/7zn/8QFxcHwNq1axk9evQTt1OvXj2srKwIDw8HIDY2lh07dtCpU6cC7bMtW7YA8Pvvv3Pw4EG6du1Kly5d+O9//6v/Zb1x40ZcXFyoU6cOvr6+fPfddyilyMnJYf369QVu63Hq1q2LnZ0dW7duBeD27dsEBgZy9uxZDh48iJ+fHyNHjqR58+bs2rWLvLy8R7ZhaWlpcCRoSKVKlfR33MTGxnL06FEAkpKS6NatGy4uLowZM4a33347311rxnTr1o0NGzYAcPfuXXbt2vXIAf5JWrZsydWrVzl9+jQAv/76Kz///DPt27enW7du/Pjjj+Tk5KDVavUjkYdFREQwZswYWrVqxVtvvcWAAQP0/XzcfjL092yMjY0Ns2fPZs2aNZw7d+6J32O3bt0IDw8nNTUVQL9/yio5ZVSCzJo1i379+rF//366dOnCypUr+eKLLwgODsbCwoK8vDz8/f1ZuXIl1tbWAEyYMAF7e3v9Bcbs7Gzat2/PF1988cj2q1SpwsKFC5k6dSp5eXk4OjqyaNEinJycmDJlCuPHj6dy5cpGTxU0b95cf3riwSkAX19fxo8fz9ixY9FoNDg6OvL5558/8YBibW3NF198wdy5c1m8eDF5eXm88cYb+Pj4EBUVZXR/3bx5k+DgYLKyspg5cyb16tWjXr16jBkzhtGjR6PT6ahcuTJLlizBwsKCmTNnMnfuXIKCgsjNzaVLly5MmDDBaDuG2NjY8MUXXzBv3jyWL1+OVqvlL3/5C23atMHFxYV33nmHoKAgtFotnTt3Jjw8/JERjL+/P5988skjo7Ineemll3j33XcJCAigZs2a+Pj4AFC5cmVef/11xowZg52dHZaWlsydO7fA233//feZOXMmQUFBuLi4UL169XyjiYetX78+30G9cePGrFu3js8++4zQ0FCysrLQaDQsWLCAunXrUqdOHa5du8aAAQNwcHCgZs2a+lNjD3Tt2pV9+/YRGBiIg4MDzs7OhIaGAtCzZ0+mTJnC3/72N/3yhv6eC6Jt27YEBQUxZ84c1q5da/B7hPs/gIYNG4adnR0NGzZ8JO6yRKMeNyYVooR7cNfMn+8GEk9v9erVNG3alFatWpGTk8PIkSN56623HnsdorAOHDhAYmIi/fv3B2Du3LnY2toW6Be9OZ05c4YTJ04watQoAP79739z6tQpPv30U/MGZiIyQhBCANCgQQNCQ0PR6XTk5ubSu3fvIkkGcP8C84oVK1i+fDk6nQ5PT898v/ZLqrp167Js2TLWr1+PRqOhWrVq+lFLWSQjBCGEEIBcVBZCCPEHSQhCCCEASQhCCCH+IAlBCCEEUMrvMrp7Nx2drvDXxF1dHUlMTDNBRCWX9Ll8kD6XD0/bZwsLDZUqGS6hUqoTgk6nniohPFi3vJE+lw/S5/LBFH2WU0ZCCCEASQhCCCH+IAlBCCEEYOKEkJaWRmBg4GPr7Z8/f57g4GACAgKYMWNGoSs+CiGEKFomSwinTp1ixIgRj9Tqf2DKlCl88MEH7NixA6UU69evN1UoQgghCsBkCWH9+vXMnj0bd3f3R+bdunWLrKwsWrZsCdx/q9X27dtNFYoQQogCMNltp/PmzTM4Ly4uDjc3N/20m5vbI2/3EkIIc4o8eYuo6JJ5XOrjW5cWdSoV+XbN8hzCn19lqJQq1JuZHnB1dXzqGNzcnJ563dJK+lw+SJ+LxvFfE7gZn0bdGoZfA2tOpuizWRJC1apViY+P108nJCQ89tSSMYmJaU/1cIabmxPx8amFXq80kz6XD9LnopObk0dNN0f+OsS7yLddUEopjh//mdzcHHx8fPWfP22fLSw0T/whbZbbTmvUqIGtrS3Hjh0DYOvWrXTt2tUcoQghRIml0WhISkogJSWZ4nh1TbEmhPHjx+tf+r1w4UIWLFhA7969ycjI0L+iTgghyjOtVsuRIwdITEwAwN8/gF69+j7VafXCMvkpoz179uj/vWzZMv2/PT092bBhg6mbF0KIUuP27VtERIRz795dbGxscHWtgqWlZbG1X6qL2wkhRFmQk5NDVNQBzpw5iZNTRYKCBlGrVp1ij0MSghBCmNmZMyc5c+YkLVq0wsenM9bWNmaJQxKCEEKYQVZWJmlpaVSp4oa3d2tq1qyFh0c1s8YkCUEIIYrZlSuX2LdvDzY2NowYMQYrKyuzJwOQhCCEEMUmPT2N/fv3cPXqZdzc3PHz64WFRckpOi0JQQghisHdu0ls2rQWrVZLx45d8PZuU6KSAUhCEEIIk8rL02JpaYWLSyU8PZvTrJkXLi5FX4eoKJSs9CSEEGWETqfj9OnjfPfdStLT09BoNHTu3K3EJgOQEYIQQhS5pKREIiLCiY29Te3azxVL2YmiIAlBCCGKiFKKY8eO8ssvR7C2tsbfvzeNGzcplrITRUESghBCFBGNRsPdu4nUrVufLl38cHCoYO6QCkUSghBCPAOtNpeffz5Co0aeuLq64e8fUKz1h4qSJAQhhHhKv/9+k4iIcJKT72FnZ4erq1upTQYgCUEIIQotJyebw4cPEB19iooVnenXbzA1a9Y2d1jPTBKCEEIU0pkzJ4mOPoW3d2vat++MtbW1uUMqEpIQhBCiALKyMklNTcXNzR1v7zbUrFkHD4+q5g6rSElCEKIYRJ68RVR0rMnbsbaxJDcnz+TtlCSm6nNMXBq13R1RSnHlyiX279+DjY3tQ8XoylYyAEkIQhSLqOhY/QFGlA613R1p1cCF7dt/5Nq1K7i5eZS4YnRFTRKCEMWktrsjU0Nam7QNNzcn4uNTTdpGSWOqPt+9m8TGjWvJy9PSsWNXvL1bl+lkAJIQhBAin4eL0TVpUrKL0RW1sp3uhBCigHQ6HadOHeO771aQnp4GQKdOXctNMgAZIQghBElJCX8Uo7tDnTp1AUpN/aGiJAlBCFFuKaX45ZcjHDsWhY2NLT179qFBg8blMhmAJAQhRDmm0WhITr5HvXoN6dLFD3t7B3OHZFaSEIQQ5Upubi4//3yYRo2a4OpaBT+/XqW6/lBRkoQghCg3bt26QUREOCkpydjb2+PqWkWSwUMkIQghyrzs7GwOH97HuXNnqFjRmf79B1OjRukvRlfUJCEIIcq8s2dPcv78Wby929C+facyU4yuqBUoIeTk5HDr1i0sLS2pVq2a7EwhRImXmZlBWlqavhhdrVrP4e7uYe6wSrQnJoQLFy7w+eefs2/fPmxtbbG0tCQnJwc/Pz9ee+01GjVqVFxxCiFEgSiluHz5Ivv3R2Bra8vIkS9jZWUlyaAADCaEL774gl9++YXBgwczb948nJ2dAUhLS+PAgQPMmzePdu3a8eabbxZbsEII8SRpaans3bub3367iru7B35+AeX2mYKnYTAhNGrUiIkTJz7yuaOjI71796Z3797s2rXriRvftm0bX375JVqtltGjRxMSEpJvfnR0NB988AG5ublUq1aNjz/+mIoVKz5lV4QQ5dn9YnRr0Ol0dOrUDS+vVmW+GF1RM7i3evToYXCly5cvG10mNjaWRYsWsWbNGrZs2cL333+vX++BefPmMWnSJH788Ufq1q3LihUrChu/EKKc02q1ALi4VKJZMy+GDRtFy5ZtJBk8BYN77ObNm7zzzjuEhoaSmZkJQHp6OgsWLGDAgAFGN3zo0CF8fHxwcXHBwcGBgIAAtm/fnm8ZnU5Heno6AJmZmdjZ2T1DV4QQ5YlOp+PEiV/47LPPSEu7X/66Y8euODu7mDewUszgKaPp06fTqFEj4uLiWLJkCf7+/kyaNAlHR0eWL19udMNxcXG4ubnpp93d3Tl9+nS+ZaZNm8bYsWOZP38+9vb2rF+/vlDBu7o+/ctG3Nycnnrd0kr6bD7WNvcffiqOeEpKn00pNjaWH3/8kd9//51GjRpRpYoTTk5lv98PM8X3bDAh3Llzh2+//ZasrCyCg4NZv349Y8aMYezYsVhZGb9bVafT5buYo5TKN52VlcWMGTP4+uuv8fLy4t///jdTp05l6dKlBQ4+MTENnU4VePkH5CUi5UNJ6vODVzyaOp6S1GdTUErx88+HOX78KLa2tvTs2ZeOHduQkJBGVlbZ7fefPe33bGGheeIPaYNHdgeH+0We7OzsSE5OJiwsDF9f3wI3XLVqVX755Rf9dHx8PO7u7vrpS5cuYWtri5eXFwDDhg3js88+K/D2hRDlj0ajITU1hQYNGtO5c3fs7e3lLqIiVKCrLpUrVy5UMgDo1KkThw8fJikpiczMTMLDw+natat+fp06dbhz5w5Xr14FYPfu3bRo0aJQbQghyr7c3FwOHtxLQkI8Sim6d+9Jjx4vYG9vb+7QyhyDI4SHs25BThH9mYeHB5MnT2bUqFHk5uYyePBgvLy8GD9+PJMmTaJFixYsWLCAt99+G6UUrq6uzJ8//+l6IYQok27ejCEycicpKclUqOAoxehMTKOUeuxJ+KZNm+rv+snKytL/+8G1gOPHjxdflAbINYSCkz6b199X3///ZWpIa5O2U5L6/Cyys7M4dGgf58+fxdnZBT+/XlSvXvOxy5aVPhdGsV9D2LlzZ6EbE0KIZ6WU4uzZU1y4EE2rVm1p164jVlZSP604GEwINWrU4NKlS1y/fh1vb288PKQOiBDCdDIy0klLS8XNzYOWLdtQu3Zd3Nzcja8oiozBi8obN27kxRdfZNmyZfTr148DBw4UZ1xCiHJCKcXFi+dYu/Ybdu78L0opLC2tJBmYgcERwqpVq9i2bRseHh6cOHGCRYsWFfpOIyGEeJLU1BT27t1FTMx1PDyq4efXS0pOmNETbx96cJqoVatW3L17t1gCEkKUD0lJiWzcuAalFJ07d6dFi5aSDMysQLedAnKrlxCiSGi1uVhZWVOpUmWaN29Js2ZeVKzobO6wBAV8MA0eTRBCCFEY94vR/cyqVStIS0tFo9HQsWMXSQYliMERwoULF2jd+n/3TGdlZdG6desS9RyCEKJ0SEiIJyJiB/HxcdStW19ODZVQBhOCsZffCCGEMUopjh49xIkTP2Nra0dAQCD16jWUMw4llMGE8Omnn/Lxxx8XZyxCiDIoLS2Vhg096dy5G3Z2Un+oJDOYEH799dfijEMIUUbk5uYQFXUQT8/mVKniJreSliIFvstICCGMuXHjNyIjd5KamoKTU0WqVHGTZFCKGEwIFy9ezHdR+QG5qCyE+LOsrCwOHdrLhQvRuLhUYuDAYVSrVsPcYYlCMpgQ6tatW6i3lwkhyielFNHRp7l48RytW7enbVufpyqZL8zP4LdmY2NDjRqS4YUQj5eRkU5qairu7h60bNmaOnXqUqWKm/EVRYllMCHY2toWZxxCiFLiQTG6gwcjsbOzZ8SIMVhaWkkyKAMMJoQvvvjC6MqJiYm4uroWaUBCiJIrNTWFyMid3LjxG1WrVpc7iMoYgwlh+vTpdOjQgeDgYJyd8z9anpaWxrp164iKimLZsmUmD1IIYX7/K0YHXbr40bx5S7kbsYx54ghh5cqVBAYGUrduXerUqYNOpyMmJoZr164xatSoAo0ihBClW25uLtbW94vRtWjRimbNvHByqmjusIQJGEwIFhYWvPLKK7z44oscOXKEq1evotFo6NmzJ506dcLGxqY44xRCFLO8vDxOnjzG6dPHGDLkRRwdnfDxkXeilGVG7w2zs7Oje/fudO/evRjCEaJkiDx5i6jo2CLbXkxcGrXdDb/cvKSJj48jImIHCQnx1KvXEAsLKX9fHsjNwkI8RlR0LBdv3KNxLZci2V5td0c6NCv57yVXShEVdZATJ37G3t6BgIAg6tdvaO6wRDGRhCCEAY1ruTA15NGn9csyjUZDRkY6jRs3pVOnbtjZ2Zk7JFGMJCEIUc7l5OQQFXWAJk2aU6WKO92795RbScspo996eno6H374IaNHj+bevXt88MEHpKenF0dsQggTi4m5xrp133DmzElu3boJIMmgHDP6zc+dO5eKFSuSmJiIra0taWlpfPDBB8URmxDCRLKyMtm9ezs//bQZKytrBg4cjrd3+To9Jh5lNCGcP3+eyZMnY2Vlhb29PQsXLuT8+fPFEZsQwkSio0/z668XaNOmA0OHvki1atXNHZIoAYxeQ/jz8DEvL0+GlEKUQunpaaSlpeLhUY2WLdvw3HP1cHWV+kPif4wmhHbt2vHxxx+TlZXF/v37Wb16Ne3bty+O2IQQRUApxYUL0Rw6tBd7eweGDx+NpaWVJAPxCKM/9d99910cHBxwcnJi0aJFNG7cmGnTphVHbEKIZ5SSksy2bRuJiAjH1bUKffoMkBG+MMjoCGHv3r288cYbvPHGG/rPtmzZwoABA4xufNu2bXz55ZdotVpGjx5NSEhIvvlXr15l9uzZJCcn4+bmxieffPJIIT0hxNNJSkpkw4bVaDQaunZ9nmbNvKQYnXgigwlhz549aLVawsLCUEqhlAJAq9WyePFiowkhNjaWRYsWsWnTJmxsbBg+fDgdOnSgQYMGwP1h7Ouvv86MGTPo2rUrCxcuZOnSpUyZMqXoeidEOfRwMbqWLdvQpIkXTk5O5g5LlAIGE8L58+c5cuQIiYmJfPvtt/9bwcqKMWPGGN3woUOH8PHxwcXFBYCAgAC2b9/Om2++CUB0dDQODg507doVgAkTJpCSkvIMXRGifMvLy2Pfvn0cPnyEoUPvF6Nr376zucMSpYjBhPDgNNHq1asfOdVTEHFxcbi5/e+ilbu7O6dPn9ZPx8TEUKVKFaZPn8758+epV68es2bNKnQ7QgiIi4slImIHiYkJNGjQGEtLKUYnCs/oNYQhQ4awc+dO/dPJeXl5xMTEMHny5Ceup9Pp8p2vVErlm9ZqtRw9epTvvvuOFi1a8Omnn/LRRx/x0UcfFTh4V9enrx7p5lb+htDS54KztrF8pvWLi1KK3bt3c+jQISpUqMCwYcPw9PQ0d1jFrqR/T6Zgij4bTQiTJ0/mxo0bxMfH07RpU06dOlWg206rVq3KL7/8op+Oj4/H3d1dP+3m5kadOnVo0aIFAIGBgUyaNKlQwScmpqHTqUKtc79tJ+LjUwu9XmkmfS6c3Jw8gFKxzxIT7+Hp2YxOnbpSs6ZbqYi5KMnfdsFZWGie+EO6QE8qb9q0ieeff57p06ezdu1akpOTjTbcqVMnDh8+TFJSEpmZmYSHh+uvFwC0atWKpKQkLly4ANy/iN2sWbOC9EmIci0nJ5t9+3YTHx8HQPfuPfHz64WtrVQmFc/G6AjB3d0dKysrnnvuOS5dusQLL7xAaqrxzOTh4cHkyZMZNWoUubm5DB48GC8vL8aPH8+kSZNo0aIF//rXv5g5cyaZmZlUrVqVsLCwIumUEGXVb79dJTJyFxkZ6bi4VMbNzV2eKxBFxmhCcHBwYNu2bXh6erJ+/Xrq1atHRkZGgTYeFBREUFBQvs+WLVum/7e3tzcbNmwoZMhClD+ZmZkcOBDBr79eoFIlVwICAqlaVeoPiaJl9KfFBx98wPnz5+ncuTMWFha8+OKLjBs3rjhiE0L84dy5M1y5com2bX0YOjREkoEwCaMjhOeee4733nsPgE8//RSAy5cvmzQoIcSjxejq1q1P5cqu5g5LlGEGRwg3b97knXfeITQ0lMzMTOD+y3I++uijApWtEEI8HaUU586dYe3ab9i16//Q6XRYWlpKMhAmZ3CEMH36dBo1akRcXBxLlizB39+fSZMmUaFCBZYvX16cMQrxzCJP3iIqOrbAy8fEpVHb/emfc3laycn3iIzcya1bN6hevSZ+fvI6S1F8DCaEO3fu8O2335KVlUVwcDDr169nzJgxjB07FisreRWzKF2iomMLdZCv7e5Ih2YeJo4qvwfF6CwsLOjWrQdNm7aQYnSiWBk8sjs4OABgZ2dHcnIyYWFh+Pr6FltgQhS12u6OTA0pea+JzM3Nwdra5o9idG1p2rQFjo7l78lbYX4FGotWrlxZkoEQRSwvL4+ffz7MqlXLSU1NQaPR0L59J0kGwmwMjhAeHqrKKSIhilZs7B0iInaQlJRIw4aeWFlZmzskIQwnhIsXL9K69f3hdVZWlv7fD4rUHT9+vHgiFKIMUUpx6NA+Tp8+joNDBfr0GcBzz9Uzd1hCAE9ICDt37izOOIQoFzQaDTk52TRp0oKOHbtga2tr7pCE0DOYEGrUqFGccQhRZmVnZ3P48H6aNWuBm5sH3bv3lLuHRIkkFweEMKHr16+wd+8uMjIycHV1xc3NQ5KBKLEkIQhhApmZGezfH8HlyxepXLkKvXv3x8OjqrnDEuKJJCEIYQLnzp3l6tVfadeuI61bt5dXWopSwehzCPHx8bz66qsEBASQkJDAuHHjiIuLK47YhChV0tJSuXPndwBatmzD0KEv0a5dR0kGotQwmhA+/PBDevToga2tLc7Oznh6ejJz5sziiE2IUkEpRXT0adau/YY9e3aglJJidKJUMpoQbt26xdChQ7GwsMDa2popU6Zw+/bt4ohNiBIvOfkuW7f+wN69u3B396Bv34Fy0ViUWkavIWg0GnQ6nX46LS0t37QQ5VVSUiI//PAdlpZWdO/ekyZNmksyEKWa0YTQq1cv3n33XVJTU1m3bh0//PADL7zwQnHEJkSJlJOTg43N/WJ0bdp0oEmT5lSoUPylsoUoakYTwoQJE9iyZQs6nY5Dhw4xbNgwhgwZUhyxCVGi5OVpOXbsKGfOnGTo0BdxcqpI27Y+5g5LiCJjNCGsW7eOwMBAeUuaKNdiY28TERFOUlIijRo1wdpaitGJssdoQoiKiuLTTz/F39+foUOH0rJly2IIS4iS4X4xur2cOnUcR0cn+vYdSJ06dc0dlhAmYTQhLFq0iOTkZH766Sfmzp1LVlYWQ4YMYfTo0cURnxBmdb8YXQ7Nm3vj4+OLjY0UoxNlV4FekOPs7MywYcN47bXXcHBwYNmyZaaOSwizyc7OIiIinLi4++9g7t69J127Pi/JQJR5RkcI586dY+PGjWzfvp2mTZvyyiuv4O/vXxyxCVHsrl27zN69u8nMzKBKFXfc3aUYnSg/jCaEiRMnMmjQIH744QeqV69eHDGJEizy5C2iomPNHUaBWNtYkpuTB8DFG/doXMvF4LIZGRns37+HK1cu4erqRp8+A3B39yimSIUoGYwmhIiICPmFJPSiomOJiUujtnvpuu++cS0XOjQzfIC/cOEs165doUOHzrRs2VbqD4lyyWBCGDFiBGvXrqV169b5EoK8QlPUdndkakhrc4dhlJubE/HxqQbnp6amkpaWSrVq1fH2bkPdug2oVKlyMUYoRMliMCF89tlnAPz000+PzFNKmS4iIUzsfjG6Uxw+vJ8KFRwZMWIMlpaWkgxEuWcwIbi7uwMwe/Zsli9fnm/e0KFDWb9+vWkjE8IE7t27S0REOLdv36JmzdryOkshHmIwIUyaNIlr165x48YNgoKC9J9rtVpsbGyKJTghilJSUgI//LAaKysr/Px64enZTJKBEA8xmBDee+89bt26xaxZs5g1a5b+c0tLSxo0aFCgjW/bto0vv/wSrVbL6NGjCQkJeexykZGRzJkzhz179hQyfCGMy87OxtbWlkqVXGnTxocmTZpJMTohHsNgQqhZsyY1a9Zkx44dT/UrKjY2lkWLFrFp0yZsbGwYPnw4HTp0eCSZJCQk8Pe//73wkQthhFar5ciRA0RHn2Lo0Jf+KEbXwdxhCVFiGXxSecSIEQC0bt0633+tWrWidWvjd5gcOnQIHx8fXFxccHBwICAggO3btz+y3MyZM3nzzTefoQtCPOr27d9ZsmQJx48f5bnn6ksxOiEK4KnuMiqIuLg43Nzc9NPu7u6cPn063zLffvstTZs2xdvb+6nacHV9+mG/m5vTU69bWhVFn61tLItsW6aglGL79u0cPXoUZ2dnQkJCCnyKs6woqd+NKUmfi4bRu4wqV67MxYsXadmyJWvXruXMmTMF+kWv0+ke+/zCA5cuXSI8PJyvv/6aO3fuPFXwiYlp6HSFvwXW2P3pZVFR9fnBk78lef+lpWXSokVL+vbtTUpKTomOtajJ33b58LR9trDQPPGHtNHidu+//z67d+/m9OnTLF++nGrVquW7yGxI1apViY+P10/Hx8frkwzA9u3biY+PZ9CgQbz66qvExcUxcuRIo9sV4s+ysrLYs2cHcXH3f1h069aDLl38sbWVYnRCFIbRhHDjxg3eeecdIiIiGDhwIG+99Rb37t0zuuFOnTpx+PBhkpKSyMzMJDw8nK5du+rnT5o0iR07drB161aWLl2Ku7s7a9aseabOiPLnypVfWbfuGy5ePKevTiq3kgrxdIwmBK1WC8CBAwfw8fEhLy+PjIwMoxv28PBg8uTJjBo1igEDBhAYGIiXlxfjx4/nzJkzzx65KNcyMtLZvn0bO3Zsw8HBgcGDQ2je/OmuRQkh7jNa3K5Vq1b06dMHS0tLWrduzejRo+nUqVOBNh4UFJTvoTbgse9SqFmzpjyDIArlwoVofvvtKj4+vnh7t5FidEIUAaMJYdasWZw4cQJPT08sLCwYN25cvlM/QhSXlJRk0tPTqFatBt7ebahXryEuLpXMHZYQZYbRhGBpaUlcXBwbN24kNzeXzp07Y2FRoBetCVEklFKcPXuSw4cP4Oj4v2J0kgyEKFpGE8KKFSv48ccfGThwIEopvv76a27fvs3EiROLIz5Rzt29m0RERDh37vxOrVp1pBidECZkNCFs2bKFtWvX4uh4/97VwYMHM3ToUEkIwuT+V4zOGn//3jRu3ESSgRAmZDQhAPpkAODk5ISVVYFWE+KpZGdnYWtrR6VKrrRr1xFPz2Y4OFQwd1hClHlGLwbUqFGDb775htzcXHJzc/n666/l3crCJLTaXA4f3s93360gNTUFjUZD69btJRkIUUyM/tT/8MMPeffddwkLCwPA29ubhQsXmjwwUTiRJ28RFR1rcP7DL5x/FqZ6n/Lt27eIiAjn3r27eHo2w9pa3rkhRHEzmhA8PDxYtWoVmZmZ6HQ6KlSQX2slUVR0rMkO1g+r7e74xJfVF5ZSigMHIjhz5iROThUJChpErVp1imz7QoiCM5gQrl+/zjvvvMO1a9fw8fEhNDQUV1fX4oxNFFJtd0emhjy+NHlJLQCm0WjQ6XS0aNEKH5/OMjIQwowMXkOYM2cOAwcO5IcffqBOnTr6U0ZCPKusrEx2795ObOz9YnRduz5Ply5+kgyEMDODI4SEhARefPFFAN5991369+9fbEGJskkpxZUrv7J//x6ys7Pw8KiGh0dVuZVUiBLCYEJ4+NZSS0tLudVUPJP09DT27dvDtWuXcXPzIChoEFWquBlfUQhRbAwe5ZXK/+IZ+RUnnsXFi+eIiblGx45d8PZuI+VPhCiBDCaEO3fuMHfuXIPTM2fONG1kotRLSUkmLS2V6tVr4u3dhvr1G+Hs7GLusIQQBhhMCCEhIU+cFsIQnU7HmTMniYo6gKOjk74YnSQDIUo2gwmhIO9NFuLPkpISiYgIJzb2NrVr16Vbtx5yulGIUkKuFIsik5h4vxidjY01PXq8QMOGnpIMhChFJCGIZ5aVlYWdnR2VK7vSoUMnGjduhoODg7nDEkIUktzqIZ7a/WJ0+/juuxWkpCSj0Who1aqdJAMhSimjCUGn07F8+XKmTp1KWloaS5YsIS/v2YukidLt1q0bfP/9Kk6c+IX69Rtia2tr7pCEEM/I6CmjsLAwkpKSOHPmDAD79+8nPj5ebjstp5RS7Nu3h+joU1Ss6Ey/foOpWbO2ucMSQhQBownh8OHDbN68meDgYBwdHVm5cqWUsSjHHlwk9vZuQ/v2nbC2tjZzREKIomI0IVhZWeV7qtTGxkbKWJQzmZmZHDwYSYsWrXB396BrV3+5e0iIMsjokb1Ro0asXr2avLw8rl69ytdff42np2dxxCbMTCnF5cuXOHBgD9nZ2VSvXgMPj6rmDksIYSJGLyrPmDGD6OhoEhMTGTFiBOnp6UyfPr04YhNmlJaWyv/934/s3PkfnJwqMmRICE2bepk7LCGECRkdITg6OjJ//vziiEWUIJcunefmzd/o1KkrXl6tpRidEOWA0YTwcEG7h8ldRmVPcvI90tPTpBidEOWU0Z99Li4u+v8qVKjA0aNHiyMuUYx0Oh0nTx7j+++/JTJyF0opKUYnRDlkdITw5yJ348eP5/XXXzdZQOVN5MlbREXHPvN2YuLSqO3uWOj1EhMTiIgIJy7uDnXq1KNbt+flDiIhyqlC3z/q6OhIXFycKWIpl6KiY7l44x6Na7k803ZquzvSoZlHoda5X4zuO2xsbOnZsw8NGjSWZCBEOWY0IYSGhuoPEkopoqOjqVevXoE2vm3bNr788ku0Wi2jR49+5J0Ku3btYvHixSilqFmzJgsWLMDZ2fkpulG6Na7lwtSQ1sXWXlZWJnZ29n8Uo/PF07Mp9vZSf0iI8s7oNYRKlSrpryFUqlSJfv368fHHHxvdcGxsLIsWLWLNmjVs2bKF77//nsuXL+vnp6Wl8be//Y2lS5fy448/0rhxYxYvXvxsvRFPlJuby8GDe/9UjK6tJAMhBFCAEUJMTAxhYWGF3vChQ4fw8fHBxcUFgICAALZv366/JpGbm8vs2bPx8Lh/mqNx48Zs27at0O2Igrl27RpbtmwlJSWZpk29sLW1M3dIQogSxmhCuHDhAkqpQp9bjouLw83NTT/t7u7O6dOn9dOVKlWiZ8+ewP16+kuXLuWll14qVBvCOKUUe/fu5ty501Ss6Ez//kOoUaOWucMSQpRARhOCm5sbffv2xdvbmwoVKug/N/Ycgk6ny5dEDCWV1NRU3njjDTw9PRk4cGBhYsfVtfB31Tzg5ub01OsWJWsbS8A08TzY5xUq2NKxY0f8/PzKXTG6kvI9Fyfpc/lgij4bTAg5OTnY2NjQqlUrWrVqVegNV61alV9++UU/HR8fj7u7e75l4uLiGDduHD4+Pk9VDiMxMQ2dThV6PTc3J+LjUwu9nink5tx/t0RRxpOZmcGBAxF4ebXGw6Mabdv64u5e8Y82soqsnZKuJH3PxUX6XD48bZ8tLDRP/CFtMCEMGzaMzZs3P/IcQkF16tSJxYsXk5SUhL29PeHh4YSGhurn5+XlMWHCBF544QUmTpz4VG2I/JRS/PrrBQ4ciCAnJ4caNWrj4VFNbiUVQhSIwYSgVOF/eT/Mw8ODyZMnM2rUKHJzcxk8eDBeXl6MHz+eSZMmcefOHc6dO0deXh47duwAoHnz5sybN++Z2i2vUlNT2bdvF7/9dg0Pj6r4+fWicuUq5g5LCFGKGEwI2dnZnDt3zmBiaNasmdGNBwUFERQUlO+zZcuWAdCiRQsuXLhQmFjFE1y+fIFbt27QuXN3WrRoKcXohBCFZjAh3Lhxg7feeuuxCUGj0bB7926TBiaMu3fvLunpadSoUQsvr9bUr9+IihXL34N9QoiiYTAhNGjQgC1bthRjKKKgdDodp04d4+jRQzg5OTNixGgsLS0lGQghnom8C7OUSUiIJyIinPj4WOrWrU/XrlKMTghRNAwmhLZt2xZnHKIAEhPj2bBhNba2tvTq1Zf69RtJMhBCFBmDCUFegFNyZGZmYm9vT+XKVfDx6YKnZ1Ps7OzNHZYQooyRW1FKsNzcXA4ciMxXjK5lyzaSDIQQJiHXEEqomzdjiIzcSUpKMs2be2NnJ8XohBCmJQmhhFFKERm5k/Pnz+Ls7MKAAUOpXr2mucMSQpQDkhBKGI1Gg5WVFa1ataNdOx+srMpXMTohhPlIQigB8nR5hIf/By+vVlStWh1fXz+5e0gIUezkorIZKaXIyEwnLi6Wq1cvc/duEoAkAyGEWZTrEULkyVtERceape28PC337t3jboYGZ3srhg59kcqVXc0SixBCQDkfIURFxxITl2aWtjMzM8nOyaZqZVv6dmkuyUAIYXbleoQAUNvdkakhrYulrbt3k8jISKdGjVrodDrS09NwcqpYLG0LIYQx5XqEUFx0Oh3Hjx9l/fpV7Nu3G6UUFhYWkgyEECVKuR8hmFpCQtwfxejiqFevIV26yB1EQoiSSRKCCd0vRrcGW1s7AgICqV+/kblDEkIIgyQhmEBmZgb29g5UrlyFjh270rhxE6k/JIQo8eQaQhHKzc1h//49fPfdCpKT76HRaPD2bi3JQAhRKsgIoYjExFxn795dpKam0KJFSxwcHMwdkhBCFIokhGeklCIiIpwLF6JxcanMwIHDqFathrnDEkKIQpOE8Iw0Gg3W1ja0adOBNm06YGUlu1QIUTrJ0espZGSks39/BN7erf8oRtddbiUVQpR6khAKQSnFxYvnOHgwEq1WS+3az1G1anVJBkKIMkESQgGlpCSzd+8ubtz4jWrVatC9e08qVaps7rCEEKLISEIooCtXLnHnzu906eJP8+beMioQQpQ5khCe4O7dJNLT06hZszbe3m1o0MATJycnc4clhBAmIQ+mPUZeXh7HjkXx/fer2L8/4qFidJIMhBBll4wQ/iQ+PpY9e8JJTIynfv1GUoxOCFFuSEJ4SELC/WJ09vYO9O7dj3r1Gpg7JCGEKDaSELj/XIGDQwVcXavQuXN3GjVqgp2dnbnDEkKIYmXSawjbtm2jT58+9OrVi9WrVz8y//z58wQHBxMQEMCMGTPQarWmDOcROqXjXvJdvvtupb4YnZdXK0kGQohyyWQJITY2lkWLFrFmzRq2bNnC999/z+XLl/MtM2XKFD744AN27NiBUor169ebKpxH/PbbNeLiYklPT6dp0+Y4OFQotraFEKIkMllCOHToED4+Pri4uODg4EBAQADbt2/Xz7916xZZWVm0bNkSgODg4HzzTeXgmduMn7OFf6w/Q3quLW5V3PD19cPa2trkbQshRElmsmsIcXFxuLm56afd3d05ffq0wflubm7ExsYWqg1XV8dCx1Wx4l0sLCxwcXbG2cWZ7q1r4eZWPm4nLS/9fJj0uXyQPhcNkyUEnU6X73ZNpVS+aWPzCyIxMQ2dThVqnRZ1KuE/sx/x8an6zx7+d1nl5uZULvr5MOlz+SB9LjgLC80Tf0ib7JRR1apViY+P10/Hx8fj7u5ucH5CQkK++UIIIYqXyRJCp06dOHz4MElJSWRmZhIeHk7Xrl3182vUqIGtrS3Hjh0DYOvWrfnmCyGEKF4mSwgeHh5MnjyZUaNGMWDAAAIDA/Hy8mL8+PGcOXMGgIULF7JgwQJ69+5NRkYGo0aNMlU4QgghjNAopQp3Er4EeZprCCDnHMsL6XP5IH0uOLNdQxBCCFG6SEIQQggBSEIQQgjxh1Jd3M7C4unLUj/LuqWV9Ll8kD6XD0/TZ2PrlOqLykIIIYqOnDISQggBSEIQQgjxB0kIQgghAEkIQggh/iAJQQghBCAJQQghxB8kIQghhAAkIQghhPiDJAQhhBBAGU8I27Zto0+fPvTq1YvVq1c/Mv/8+fMEBwcTEBDAjBkz0Gq1ZoiyaBnr865du+jfvz/9+vVj4sSJJCcnmyHKomWszw9ERkbi7+9fjJGZjrE+X716lZdeeol+/foxbty4cvE9R0dHM2jQIPr168drr71GSkqKGaIsWmlpaQQGBnLz5s1H5pnk+KXKqDt37ig/Pz919+5dlZ6eroKCgtSvv/6ab5m+ffuqEydOKKWUev/999Xq1avNEGnRMdbn1NRU1blzZ3Xnzh2llFKffvqpCg0NNVe4RaIg37NSSsXHx6vevXsrPz8/M0RZtIz1WafTqV69eqm9e/cqpZT6+OOPVVhYmLnCLRIF+Z5HjBihIiMjlVJKLViwQH3yySfmCLXInDx5UgUGBqpmzZqpGzduPDLfFMevMjtCOHToED4+Pri4uODg4EBAQADbt2/Xz7916xZZWVm0bNkSgODg4HzzSyNjfc7NzWX27Nl4eHgA0LhxY27fvm2ucIuEsT4/MHPmTN58800zRFj0jPU5OjoaBwcH/StpJ0yYQEhIiLnCLRIF+Z51Oh3p6ekAZGZmYmdnZ45Qi8z69euZPXv2Y981b6rjV5lNCHFxcbi5uemn3d3diY2NNTjfzc0t3/zSyFifK1WqRM+ePQHIyspi6dKl9OjRo9jjLErG+gzw7bff0rRpU7y9vYs7PJMw1ueYmBiqVKnC9OnTGThwILNnz8bBwcEcoRaZgnzP06ZNY+bMmfj6+nLo0CGGDx9e3GEWqXnz5tG2bdvHzjPV8avMJgSdTodG879Sr0qpfNPG5pdGBe1Tamoqr776Kp6engwcOLA4Qyxyxvp86dIlwsPDmThxojnCMwljfdZqtRw9epQRI0awefNmatWqxUcffWSOUIuMsT5nZWUxY8YMvv76aw4cOMDIkSOZOnWqOUItFqY6fpXZhFC1alXi4+P10/Hx8fmGXn+en5CQ8NihWWlirM9w/5fFyJEjady4MfPmzSvuEIucsT5v376d+Ph4Bg0axKuvvqrvf2lmrM9ubm7UqVOHFi1aABAYGMjp06eLPc6iZKzPly5dwtbWFi8vLwCGDRvG0aNHiz3O4mKq41eZTQidOnXi8OHDJCUlkZmZSXh4uP6cKkCNGjWwtbXl2LFjAGzdujXf/NLIWJ/z8vKYMGECL7zwAjNmzCj1IyIw3udJkyaxY8cOtm7dytKlS3F3d2fNmjVmjPjZGetzq1atSEpK4sKFCwDs2bOHZs2amSvcImGsz3Xq1OHOnTtcvXoVgN27d+sTYllksuPXM1+WLsF+/PFH1bdvX9WrVy+1dOlSpZRSr7zyijp9+rRSSqnz58+rQYMGqYCAAPXXv/5VZWdnmzPcIvGkPoeHh6vGjRurfv366f+bPn26mSN+dsa+5wdu3LhRJu4yUsp4n0+ePKkGDRqk+vTpo8aOHasSEhLMGW6RMNbnyMhIFRQUpAIDA9Xo0aNVTEyMOcMtMn5+fvq7jEx9/JI3pgkhhADK8CkjIYQQhSMJQQghBCAJQQghxB8kIQghhAAkIQghhPiDlbkDEKVf48aNadSoERYW//t90bx58yc++LZp0yZ27NjBkiVLnrn9xYsXs3r1ajw8PNBoNOTl5eHq6srs2bOpW7duobcXGxvLX/7yF9atW8eNGzcICwtj8eLF+T5/Vjdv3qRnz540atRI/1lGRgZVq1Zl/vz51KpV64nrf/7553h6eha69EheXh4TJ05k3rx5XLlyhfHjxz+yj4YPH06XLl0eiU8pxahRoxg8eHCB4t+1axcXL17kjTfeKFSMwnwkIYgi8c0331C5cmWztd+nTx8++OAD/fSqVat455132LRpU6G35eHhoT/o//7771y7du2Rz4uCnZ0dW7du1U8rpZg7dy6LFi3ik08+eeK6UVFRNGjQoNBtrly5kvbt21OlShWuXLlC7dq188XwwM2bNx+JLzY2lsDAQJo3b46jo6PR+Hv06MHq1as5f/48TZo0KXSsovjJKSNhUhs2bGDIkCEMGDAAPz+/xz4lHB4ezsCBAwkODmbIkCH8/PPPwP2aS9OmTSM4OJigoCDmz59f4JrvHTt21B/I79y5w4QJEwgKCiIwMJDly5cD92v+zJ49m6CgIIKDg5k0aRLp6encvHmTVq1akZeXx8yZM4mJiWHcuHH5Pu/WrRtnz57Vt/f222/r+/bll18ycOBA+vfvz8SJEwtcdCw7O5u4uDicnZ0BuHbtGi+//DJDhw7Fz8+P119/nezsbFavXs3Zs2cJCwtj586d5OTkMH/+fAYOHEi/fv2YNm0aaWlpj2w/MzOTb775huDg4ALF82ceHh7UqVOH69evFyh+gMGDB/P5558/VXui+MkIQRSJ0aNH5ztltHLlSuzs7Pjhhx9YunQplSpV4uTJk7z88suP1BIKCwtj4cKFtGzZkgMHDhAVFUW7du2YP38+zZo146OPPiIvL49p06bx73//m/Hjxz8xFq1Wy4YNG+jQoQMA7777Ls8//zwvv/wyqamphISEUK1aNTw8PDh69Cj//e9/0Wg0fPzxx1y8eFFfE8bS0pK5c+cSGhrKihUr9C8psbS0ZNCgQWzatInmzZuTnJzM4cOHCQ0NZcuWLVy6dIkffvgBKysrvv/+e2bOnMmyZcseiTMrK4v+/fuj0+lITEzE2dmZXr168eqrrwL3yx8PGDCA/v37k5ubS3BwMJGRkYSEhLB9+3ZCQkLo2bMnn3/+OZaWlmzatAmNRsMnn3zCwoUL+dvf/pavvSNHjlC3bl0qVaqk/ywmJob+/fvrp6tVq8ZXX3312P164sQJYmJi8Pb2Ji8vz2j8AJ07d2batGlkZWWV+nLU5YEkBFEkDJ0y+uqrr9i7dy/Xr1/nwoULZGRkPLJM3759efPNN+nWrRudO3fWH/AjIyM5c+YMGzZsAO4fQA3573//q6/rkpubS7NmzQgNDSUjI4Pjx4+zcuVKAJycnAgODmbfvn3MmDEDS0tLhgwZgq+vLwEBAXh5eT327VR/NmjQIAYPHsy0adP46aef8Pf3x8nJiYiICM6cOcOgQYOA+1UpMzMzH7uNh0+57N+/nylTpuDn50eFChUAmDJlCgcPHmTZsmVcv36duLi4x+6/yMhIUlNTOXTokL7/rq6ujyx39epVateune8zQ6eM4H8JC+5fe6hUqRIff/wx1apVe+SU0uPiB3BxccHW1pZbt25Rv359wztUlAiSEITJ3Llzh2HDhjF06FDatGlD7969iYiIeGS5yZMnM2jQIA4ePMimTZtYuXIlGzZsQKfT8dlnn+kPJCkpKQYL8v35GsIDaWlp/Lk6i06nQ6vVUrFiRbZu3crx48c5cuQIb7/9NuPGjaNbt25G+1ajRg2aNm1KZGQkmzZtYvr06fptv/LKK/pRUE5OToFeX9mlSxdefvll/vKXv/Cf//wHR0dH/vrXv5KXl8cLL7xA9+7duX379iN9edDm9OnT9XGnp6eTnZ39yHIajQadTmc0lgf+fI2gsPE/YGlpiaWlZYHbFeYj1xCEyZw9e5bKlSszceJEfH199ckgLy9Pv4xWq8Xf35/MzExGjBjB7NmzuXjxIjk5Ofj6+vL111+jlCInJ4fXX3+d7777rlAxODo64u3trX8Hb2pqKlu2bKFTp05EREQwZswYWrVqxVtvvcWAAQPyXReA+wez3Nzcx2576NChLFu2jMzMTNq0aQOAr68vGzZs0J/D/+yzz3jvvfcKFOvYsWOpUKEC//znPwE4cOAAb7zxBn369AHg1KlT+n1naWmpv57i6+vL6tWrycnJQafTMWvWrMdelK5bty43btwoUCxP48/xw/39nZOTQ/Xq1U3Wrig6MkIQJtO5c2c2bNhA79690Wg0tG/fnsqVK/Pbb7/pl7GysmL69Om8++67WFlZodFomD9/PjY2NsyYMYN58+YRFBREbm4unTp14pVXXil0HAsXLmTOnDls2rSJnJwc/UVknU7Hvn37CAwMxMHBAWdnZ0JDQ/Ot26BBA2xtbRk8eDCLFi3KN8/f358PP/ww3zWNIUOGEBsby9ChQ9FoNFSrVq3AL6extrZm1qxZvPLKKwwePJjJkyfzxhtv4ODggKOjI+3atSMmJkbf9ieffEJubi4TJ07k73//OwMHDiQvL48mTZowbdq0R7bfqVMnZsyYQUpKChUrVizsbix0/I0aNeLAgQN0794dGxubIm9PFD2pdipEOfLVV19haWlp9MJ8URk1ahTTp0/H09OzWNoTz0ZOGQlRjowdO5YjR47ke9uWqezcuZO2bdtKMihFZIQghBACkBGCEEKIP0hCEEIIAUhCEEII8QdJCEIIIQBJCEIIIf4gCUEIIQQA/w/wPM4R6t61IQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ROC curve\n",
    "\n",
    "fpr, tpr, threshold = met.roc_curve(ytest, logit.predict_proba(Xtest)[:,1])\n",
    "\n",
    "x = np.arange(0,1,0.01)\n",
    "\n",
    "plt.plot(x , x, '--', color = 'gray')\n",
    "plt.plot(fpr, tpr)\n",
    "plt.title(\"ROC curve for poor health using Logistic Reg\")\n",
    "plt.xlabel(\"False Positive Rate (FPR)\")\n",
    "plt.ylabel(\"True Positive Rate (TPR)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Method 2 - Impute mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>educ</th>\n",
       "      <th>married</th>\n",
       "      <th>income</th>\n",
       "      <th>health_excellent</th>\n",
       "      <th>health_fair</th>\n",
       "      <th>health_good</th>\n",
       "      <th>health_poor</th>\n",
       "      <th>sex_female</th>\n",
       "      <th>sex_male</th>\n",
       "      <th>sexornt_bisexual</th>\n",
       "      <th>sexornt_dont know</th>\n",
       "      <th>sexornt_heterosexual or straight</th>\n",
       "      <th>sexornt_homosexual or gay</th>\n",
       "      <th>partyid_dem</th>\n",
       "      <th>partyid_other</th>\n",
       "      <th>partyid_rep</th>\n",
       "      <th>race_black</th>\n",
       "      <th>race_other</th>\n",
       "      <th>race_white</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>53</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>26</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>59</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>74</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>37</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  educ  married   income  health_excellent  health_fair  health_good  \\\n",
       "0   53    16        1  18750.0                 1            0            0   \n",
       "1   26    16        1  18750.0                 1            0            0   \n",
       "2   59    13        1      NaN                 0            0            1   \n",
       "3   74    17        1      NaN                 1            0            0   \n",
       "4   37    10        0      NaN                 0            0            0   \n",
       "\n",
       "   health_poor  sex_female  sex_male  sexornt_bisexual  sexornt_dont know  \\\n",
       "0            0           0         1                 0                  0   \n",
       "1            0           1         0                 0                  0   \n",
       "2            0           0         1                 0                  0   \n",
       "3            0           1         0                 0                  0   \n",
       "4            1           1         0                 0                  0   \n",
       "\n",
       "   sexornt_heterosexual or straight  sexornt_homosexual or gay  partyid_dem  \\\n",
       "0                                 1                          0            0   \n",
       "1                                 1                          0            0   \n",
       "2                                 1                          0            0   \n",
       "3                                 1                          0            0   \n",
       "4                                 1                          0            0   \n",
       "\n",
       "   partyid_other  partyid_rep  race_black  race_other  race_white  \n",
       "0              0            1           0           0           1  \n",
       "1              0            1           0           0           1  \n",
       "2              0            1           0           0           1  \n",
       "3              1            0           0           0           1  \n",
       "4              1            0           0           0           1  "
      ]
     },
     "execution_count": 224,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get the original data and \n",
    "# Convert categorical variables into dummy variables\n",
    "gssdata2 = pd.read_csv('./data/gssdata4.csv')\n",
    "gssdata2 = pd.get_dummies(gssdata2)\n",
    "gssdata2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fill missing income values with mean\n",
    "gssdata2['income'] = gssdata2.fillna(np.mean(gssdata2['income']))\n",
    "\n",
    "# check that there are no more null values\n",
    "sum(gssdata2.income.isnull())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score =  0.9368556701030928\n",
      "coeffients betas \n",
      " [[ 0.01076633  0.01076633 -0.20084665 -0.36431352  0.05297761 -0.907603\n",
      "   0.01076633]]\n"
     ]
    }
   ],
   "source": [
    "# Specify the X and Y vectors\n",
    "X2 = gssdata2[['income','age','educ','sex_female','partyid_dem','partyid_rep','income']]\n",
    "Y2 = gssdata2['health_poor']\n",
    "\n",
    "# split into train and test\n",
    "from sklearn.model_selection import train_test_split\n",
    "Xtrain2, Xtest2, ytrain2, ytest2 = train_test_split(X2, Y2, random_state=1, train_size = 0.5)\n",
    "\n",
    "# linear models from sklearn already imported as lm\n",
    "\n",
    "logit2 = lm.LogisticRegression()\n",
    "logit2.fit(Xtrain2, ytrain2)\n",
    "print(\"score = \", logit2.score(Xtest2, ytest2))\n",
    "print(\"coeffients betas \\n\", logit2.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[727   1]\n",
      " [ 48   0]]\n",
      "[[717  11]\n",
      " [ 46   2]]\n",
      "[[597 131]\n",
      " [ 29  19]]\n"
     ]
    }
   ],
   "source": [
    "# confusion_matrix(actual, pred)\n",
    "# Columns are for predicted (0,1) and rows are observed (0,1)\n",
    "ypred2 = logit2.predict(Xtest2)\n",
    "yhat2 = logit2.predict_proba(Xtest2) # probabilities\n",
    "\n",
    "# As you reduce the probability threshold to classify as 1, number of 1s go up\n",
    "print(confusion_matrix(ytest2, ypred2))\n",
    "print(confusion_matrix(ytest2, yhat2[:,1] > 0.3))\n",
    "print(confusion_matrix(ytest2, yhat2[:,1] > 0.1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'True Positive Rate (TPR)')"
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ROC curve\n",
    "\n",
    "fpr2, tpr2, threshold2 = met.roc_curve(ytest2, logit2.predict_proba(Xtest2)[:,1])\n",
    "\n",
    "x = np.arange(0,1,0.01)\n",
    "\n",
    "plt.plot(x , x, '--', color = 'gray')\n",
    "plt.plot(fpr2, tpr2)\n",
    "plt.title(\"ROC curve for poor health using Logistic Reg\")\n",
    "plt.xlabel(\"False Positive Rate (FPR)\")\n",
    "plt.ylabel(\"True Positive Rate (TPR)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Method 3 - Impute NAs using regression\n",
    "Build a regression with income = f(other variables) and then predict the missing values of income using this function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>educ</th>\n",
       "      <th>married</th>\n",
       "      <th>income</th>\n",
       "      <th>health_excellent</th>\n",
       "      <th>health_fair</th>\n",
       "      <th>health_good</th>\n",
       "      <th>health_poor</th>\n",
       "      <th>sex_female</th>\n",
       "      <th>sex_male</th>\n",
       "      <th>sexornt_bisexual</th>\n",
       "      <th>sexornt_dont know</th>\n",
       "      <th>sexornt_heterosexual or straight</th>\n",
       "      <th>sexornt_homosexual or gay</th>\n",
       "      <th>partyid_dem</th>\n",
       "      <th>partyid_other</th>\n",
       "      <th>partyid_rep</th>\n",
       "      <th>race_black</th>\n",
       "      <th>race_other</th>\n",
       "      <th>race_white</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>53</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>26</td>\n",
       "      <td>16</td>\n",
       "      <td>1</td>\n",
       "      <td>18750.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>59</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>74</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>37</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  educ  married   income  health_excellent  health_fair  health_good  \\\n",
       "0   53    16        1  18750.0                 1            0            0   \n",
       "1   26    16        1  18750.0                 1            0            0   \n",
       "2   59    13        1      NaN                 0            0            1   \n",
       "3   74    17        1      NaN                 1            0            0   \n",
       "4   37    10        0      NaN                 0            0            0   \n",
       "\n",
       "   health_poor  sex_female  sex_male  sexornt_bisexual  sexornt_dont know  \\\n",
       "0            0           0         1                 0                  0   \n",
       "1            0           1         0                 0                  0   \n",
       "2            0           0         1                 0                  0   \n",
       "3            0           1         0                 0                  0   \n",
       "4            1           1         0                 0                  0   \n",
       "\n",
       "   sexornt_heterosexual or straight  sexornt_homosexual or gay  partyid_dem  \\\n",
       "0                                 1                          0            0   \n",
       "1                                 1                          0            0   \n",
       "2                                 1                          0            0   \n",
       "3                                 1                          0            0   \n",
       "4                                 1                          0            0   \n",
       "\n",
       "   partyid_other  partyid_rep  race_black  race_other  race_white  \n",
       "0              0            1           0           0           1  \n",
       "1              0            1           0           0           1  \n",
       "2              0            1           0           0           1  \n",
       "3              1            0           0           0           1  \n",
       "4              1            0           0           0           1  "
      ]
     },
     "execution_count": 246,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get the original data and \n",
    "# Convert categorical variables into dummy variables\n",
    "gssdata3 = pd.read_csv('./data/gssdata4.csv')\n",
    "gssdata3 = pd.get_dummies(gssdata2)\n",
    "gssdata3.head()"
   ]
  }
 ],
 "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
}