case2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MSDS 7331 - Case Study 2 - Diabetes Prediction\n",
    "Daniel Crouthamel\n",
    "\n",
    "Sophia Wu\n",
    "\n",
    "Fabio Savorgnan\n",
    "\n",
    "Bo Yun\n",
    "\n",
    "# Introduction\n",
    "\n",
    "In this study, we will be building a classifier to predict less than 30 days readmission of patients with Diabetes\n",
    "\n",
    "Business Understanding\n",
    "\n",
    "Data Evaluation / Engineering\n",
    "\n",
    "Modeling Preparations\n",
    "\n",
    "Model Building & Evaluation\n",
    "\n",
    "Model Interpretability & Explainability\n",
    "\n",
    "Case Conclusions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.model_selection import ShuffleSplit\n",
    "from sklearn import metrics as mt\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import roc_curve\n",
    "from matplotlib import pyplot\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.calibration import CalibratedClassifierCV\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "from sklearn.metrics import classification_report,confusion_matrix\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.compose import make_column_selector as selector\n",
    "from sklearn.preprocessing import RobustScaler\n",
    "# Imbalanced Data\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from imblearn.pipeline import Pipeline as imbpipeline\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV, StratifiedKFold\n",
    "\n",
    "# Pre-Processing\n",
    "from sklearn.preprocessing import RobustScaler\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.compose import ColumnTransformer\n",
    "\n",
    "from sklearn.metrics import make_scorer\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import recall_score\n",
    "from sklearn.metrics import precision_score\n",
    "from sklearn.metrics import f1_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Bussiness Understanding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "According to the Healthcare Cost and Utilization Project (H-CUP) conducted by the Agency for Healthcare Research and Quality (AHRQ), the national benchmark 30-day all cause readmission rate for patients with diabetes is 19.8%.\n",
    "We will be building a classifier to predict the readmission in less than 30 days of being d/c from the hospital with this given databaset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Engeniering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load the data"
   ]
  },
  {
   "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>encounter_id</th>\n",
       "      <th>patient_nbr</th>\n",
       "      <th>race</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>weight</th>\n",
       "      <th>admission_type_id</th>\n",
       "      <th>discharge_disposition_id</th>\n",
       "      <th>admission_source_id</th>\n",
       "      <th>time_in_hospital</th>\n",
       "      <th>...</th>\n",
       "      <th>citoglipton</th>\n",
       "      <th>insulin</th>\n",
       "      <th>glyburide-metformin</th>\n",
       "      <th>glipizide-metformin</th>\n",
       "      <th>glimepiride-pioglitazone</th>\n",
       "      <th>metformin-rosiglitazone</th>\n",
       "      <th>metformin-pioglitazone</th>\n",
       "      <th>change</th>\n",
       "      <th>diabetesMed</th>\n",
       "      <th>readmitted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2278392</td>\n",
       "      <td>8222157</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Female</td>\n",
       "      <td>[0-10)</td>\n",
       "      <td>?</td>\n",
       "      <td>6</td>\n",
       "      <td>25</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>149190</td>\n",
       "      <td>55629189</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Female</td>\n",
       "      <td>[10-20)</td>\n",
       "      <td>?</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>Up</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Ch</td>\n",
       "      <td>Yes</td>\n",
       "      <td>&gt;30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>64410</td>\n",
       "      <td>86047875</td>\n",
       "      <td>AfricanAmerican</td>\n",
       "      <td>Female</td>\n",
       "      <td>[20-30)</td>\n",
       "      <td>?</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>500364</td>\n",
       "      <td>82442376</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>[30-40)</td>\n",
       "      <td>?</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>Up</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Ch</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>16680</td>\n",
       "      <td>42519267</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>[40-50)</td>\n",
       "      <td>?</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>Steady</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Ch</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   encounter_id  patient_nbr             race  gender      age weight  \\\n",
       "0       2278392      8222157        Caucasian  Female   [0-10)      ?   \n",
       "1        149190     55629189        Caucasian  Female  [10-20)      ?   \n",
       "2         64410     86047875  AfricanAmerican  Female  [20-30)      ?   \n",
       "3        500364     82442376        Caucasian    Male  [30-40)      ?   \n",
       "4         16680     42519267        Caucasian    Male  [40-50)      ?   \n",
       "\n",
       "   admission_type_id  discharge_disposition_id  admission_source_id  \\\n",
       "0                  6                        25                    1   \n",
       "1                  1                         1                    7   \n",
       "2                  1                         1                    7   \n",
       "3                  1                         1                    7   \n",
       "4                  1                         1                    7   \n",
       "\n",
       "   time_in_hospital  ... citoglipton insulin  glyburide-metformin  \\\n",
       "0                 1  ...          No      No                   No   \n",
       "1                 3  ...          No      Up                   No   \n",
       "2                 2  ...          No      No                   No   \n",
       "3                 2  ...          No      Up                   No   \n",
       "4                 1  ...          No  Steady                   No   \n",
       "\n",
       "   glipizide-metformin  glimepiride-pioglitazone  metformin-rosiglitazone  \\\n",
       "0                   No                        No                       No   \n",
       "1                   No                        No                       No   \n",
       "2                   No                        No                       No   \n",
       "3                   No                        No                       No   \n",
       "4                   No                        No                       No   \n",
       "\n",
       "   metformin-pioglitazone  change diabetesMed readmitted  \n",
       "0                      No      No          No         NO  \n",
       "1                      No      Ch         Yes        >30  \n",
       "2                      No      No         Yes         NO  \n",
       "3                      No      Ch         Yes         NO  \n",
       "4                      No      Ch         Yes         NO  \n",
       "\n",
       "[5 rows x 50 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"data/diabetic_data.csv\")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Get some info from the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 101766 entries, 0 to 101765\n",
      "Data columns (total 50 columns):\n",
      " #   Column                    Non-Null Count   Dtype \n",
      "---  ------                    --------------   ----- \n",
      " 0   encounter_id              101766 non-null  int64 \n",
      " 1   patient_nbr               101766 non-null  int64 \n",
      " 2   race                      101766 non-null  object\n",
      " 3   gender                    101766 non-null  object\n",
      " 4   age                       101766 non-null  object\n",
      " 5   weight                    101766 non-null  object\n",
      " 6   admission_type_id         101766 non-null  int64 \n",
      " 7   discharge_disposition_id  101766 non-null  int64 \n",
      " 8   admission_source_id       101766 non-null  int64 \n",
      " 9   time_in_hospital          101766 non-null  int64 \n",
      " 10  payer_code                101766 non-null  object\n",
      " 11  medical_specialty         101766 non-null  object\n",
      " 12  num_lab_procedures        101766 non-null  int64 \n",
      " 13  num_procedures            101766 non-null  int64 \n",
      " 14  num_medications           101766 non-null  int64 \n",
      " 15  number_outpatient         101766 non-null  int64 \n",
      " 16  number_emergency          101766 non-null  int64 \n",
      " 17  number_inpatient          101766 non-null  int64 \n",
      " 18  diag_1                    101766 non-null  object\n",
      " 19  diag_2                    101766 non-null  object\n",
      " 20  diag_3                    101766 non-null  object\n",
      " 21  number_diagnoses          101766 non-null  int64 \n",
      " 22  max_glu_serum             101766 non-null  object\n",
      " 23  A1Cresult                 101766 non-null  object\n",
      " 24  metformin                 101766 non-null  object\n",
      " 25  repaglinide               101766 non-null  object\n",
      " 26  nateglinide               101766 non-null  object\n",
      " 27  chlorpropamide            101766 non-null  object\n",
      " 28  glimepiride               101766 non-null  object\n",
      " 29  acetohexamide             101766 non-null  object\n",
      " 30  glipizide                 101766 non-null  object\n",
      " 31  glyburide                 101766 non-null  object\n",
      " 32  tolbutamide               101766 non-null  object\n",
      " 33  pioglitazone              101766 non-null  object\n",
      " 34  rosiglitazone             101766 non-null  object\n",
      " 35  acarbose                  101766 non-null  object\n",
      " 36  miglitol                  101766 non-null  object\n",
      " 37  troglitazone              101766 non-null  object\n",
      " 38  tolazamide                101766 non-null  object\n",
      " 39  examide                   101766 non-null  object\n",
      " 40  citoglipton               101766 non-null  object\n",
      " 41  insulin                   101766 non-null  object\n",
      " 42  glyburide-metformin       101766 non-null  object\n",
      " 43  glipizide-metformin       101766 non-null  object\n",
      " 44  glimepiride-pioglitazone  101766 non-null  object\n",
      " 45  metformin-rosiglitazone   101766 non-null  object\n",
      " 46  metformin-pioglitazone    101766 non-null  object\n",
      " 47  change                    101766 non-null  object\n",
      " 48  diabetesMed               101766 non-null  object\n",
      " 49  readmitted                101766 non-null  object\n",
      "dtypes: int64(13), object(37)\n",
      "memory usage: 38.8+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(101766, 50)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Look at the hist from the numerical variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<AxesSubplot:title={'center':'encounter_id'}>,\n",
       "        <AxesSubplot:title={'center':'patient_nbr'}>,\n",
       "        <AxesSubplot:title={'center':'admission_type_id'}>,\n",
       "        <AxesSubplot:title={'center':'discharge_disposition_id'}>],\n",
       "       [<AxesSubplot:title={'center':'admission_source_id'}>,\n",
       "        <AxesSubplot:title={'center':'time_in_hospital'}>,\n",
       "        <AxesSubplot:title={'center':'num_lab_procedures'}>,\n",
       "        <AxesSubplot:title={'center':'num_procedures'}>],\n",
       "       [<AxesSubplot:title={'center':'num_medications'}>,\n",
       "        <AxesSubplot:title={'center':'number_outpatient'}>,\n",
       "        <AxesSubplot:title={'center':'number_emergency'}>,\n",
       "        <AxesSubplot:title={'center':'number_inpatient'}>],\n",
       "       [<AxesSubplot:title={'center':'number_diagnoses'}>,\n",
       "        <AxesSubplot:>, <AxesSubplot:>, <AxesSubplot:>]], dtype=object)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x1008 with 16 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.hist(figsize=(14,14), xrot=50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Get detail imformation from each column so we can replace miss values with imputation and or delated columns or row as appropriate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "race\n",
      "Caucasian          76099\n",
      "AfricanAmerican    19210\n",
      "?                   2273\n",
      "Hispanic            2037\n",
      "Other               1506\n",
      "Asian                641\n",
      "Name: race, dtype: int64\n",
      "-------------\n",
      "gender\n",
      "Female             54708\n",
      "Male               47055\n",
      "Unknown/Invalid        3\n",
      "Name: gender, dtype: int64\n",
      "-------------\n",
      "age\n",
      "[70-80)     26068\n",
      "[60-70)     22483\n",
      "[50-60)     17256\n",
      "[80-90)     17197\n",
      "[40-50)      9685\n",
      "[30-40)      3775\n",
      "[90-100)     2793\n",
      "[20-30)      1657\n",
      "[10-20)       691\n",
      "[0-10)        161\n",
      "Name: age, dtype: int64\n",
      "-------------\n",
      "weight\n",
      "?            98569\n",
      "[75-100)      1336\n",
      "[50-75)        897\n",
      "[100-125)      625\n",
      "[125-150)      145\n",
      "[25-50)         97\n",
      "[0-25)          48\n",
      "[150-175)       35\n",
      "[175-200)       11\n",
      ">200             3\n",
      "Name: weight, dtype: int64\n",
      "-------------\n",
      "admission_type_id\n",
      "1    53990\n",
      "3    18869\n",
      "2    18480\n",
      "6     5291\n",
      "5     4785\n",
      "8      320\n",
      "7       21\n",
      "4       10\n",
      "Name: admission_type_id, dtype: int64\n",
      "-------------\n",
      "discharge_disposition_id\n",
      "1     60234\n",
      "3     13954\n",
      "6     12902\n",
      "18     3691\n",
      "2      2128\n",
      "22     1993\n",
      "11     1642\n",
      "5      1184\n",
      "25      989\n",
      "4       815\n",
      "7       623\n",
      "23      412\n",
      "13      399\n",
      "14      372\n",
      "28      139\n",
      "8       108\n",
      "15       63\n",
      "24       48\n",
      "9        21\n",
      "17       14\n",
      "16       11\n",
      "19        8\n",
      "10        6\n",
      "27        5\n",
      "12        3\n",
      "20        2\n",
      "Name: discharge_disposition_id, dtype: int64\n",
      "-------------\n",
      "admission_source_id\n",
      "7     57494\n",
      "1     29565\n",
      "17     6781\n",
      "4      3187\n",
      "6      2264\n",
      "2      1104\n",
      "5       855\n",
      "3       187\n",
      "20      161\n",
      "9       125\n",
      "8        16\n",
      "22       12\n",
      "10        8\n",
      "11        2\n",
      "14        2\n",
      "25        2\n",
      "13        1\n",
      "Name: admission_source_id, dtype: int64\n",
      "-------------\n",
      "time_in_hospital\n",
      "3     17756\n",
      "2     17224\n",
      "1     14208\n",
      "4     13924\n",
      "5      9966\n",
      "6      7539\n",
      "7      5859\n",
      "8      4391\n",
      "9      3002\n",
      "10     2342\n",
      "11     1855\n",
      "12     1448\n",
      "13     1210\n",
      "14     1042\n",
      "Name: time_in_hospital, dtype: int64\n",
      "-------------\n",
      "payer_code\n",
      "?     40256\n",
      "MC    32439\n",
      "HM     6274\n",
      "SP     5007\n",
      "BC     4655\n",
      "MD     3532\n",
      "CP     2533\n",
      "UN     2448\n",
      "CM     1937\n",
      "OG     1033\n",
      "PO      592\n",
      "DM      549\n",
      "CH      146\n",
      "WC      135\n",
      "OT       95\n",
      "MP       79\n",
      "SI       55\n",
      "FR        1\n",
      "Name: payer_code, dtype: int64\n",
      "-------------\n",
      "medical_specialty\n",
      "?                                   49949\n",
      "InternalMedicine                    14635\n",
      "Emergency/Trauma                     7565\n",
      "Family/GeneralPractice               7440\n",
      "Cardiology                           5352\n",
      "                                    ...  \n",
      "Surgery-PlasticwithinHeadandNeck        1\n",
      "Pediatrics-InfectiousDiseases           1\n",
      "Psychiatry-Addictive                    1\n",
      "Perinatology                            1\n",
      "Speech                                  1\n",
      "Name: medical_specialty, Length: 73, dtype: int64\n",
      "-------------\n",
      "num_lab_procedures\n",
      "1      3208\n",
      "43     2804\n",
      "44     2496\n",
      "45     2376\n",
      "38     2213\n",
      "       ... \n",
      "107       1\n",
      "118       1\n",
      "129       1\n",
      "120       1\n",
      "121       1\n",
      "Name: num_lab_procedures, Length: 118, dtype: int64\n",
      "-------------\n",
      "num_procedures\n",
      "0    46652\n",
      "1    20742\n",
      "2    12717\n",
      "3     9443\n",
      "6     4954\n",
      "4     4180\n",
      "5     3078\n",
      "Name: num_procedures, dtype: int64\n",
      "-------------\n",
      "num_medications\n",
      "13    6086\n",
      "12    6004\n",
      "11    5795\n",
      "15    5792\n",
      "14    5707\n",
      "      ... \n",
      "70       2\n",
      "75       2\n",
      "74       1\n",
      "79       1\n",
      "81       1\n",
      "Name: num_medications, Length: 75, dtype: int64\n",
      "-------------\n",
      "number_outpatient\n",
      "0     85027\n",
      "1      8547\n",
      "2      3594\n",
      "3      2042\n",
      "4      1099\n",
      "5       533\n",
      "6       303\n",
      "7       155\n",
      "8        98\n",
      "9        83\n",
      "10       57\n",
      "11       42\n",
      "13       31\n",
      "12       30\n",
      "14       28\n",
      "15       20\n",
      "16       15\n",
      "17        8\n",
      "21        7\n",
      "20        7\n",
      "22        5\n",
      "18        5\n",
      "19        3\n",
      "24        3\n",
      "27        3\n",
      "23        2\n",
      "25        2\n",
      "26        2\n",
      "29        2\n",
      "33        2\n",
      "35        2\n",
      "36        2\n",
      "40        1\n",
      "28        1\n",
      "34        1\n",
      "37        1\n",
      "38        1\n",
      "39        1\n",
      "42        1\n",
      "Name: number_outpatient, dtype: int64\n",
      "-------------\n",
      "number_emergency\n",
      "0     90383\n",
      "1      7677\n",
      "2      2042\n",
      "3       725\n",
      "4       374\n",
      "5       192\n",
      "6        94\n",
      "7        73\n",
      "8        50\n",
      "10       34\n",
      "9        33\n",
      "11       23\n",
      "13       12\n",
      "12       10\n",
      "22        6\n",
      "18        5\n",
      "16        5\n",
      "19        4\n",
      "20        4\n",
      "14        3\n",
      "15        3\n",
      "21        2\n",
      "25        2\n",
      "76        1\n",
      "54        1\n",
      "24        1\n",
      "28        1\n",
      "29        1\n",
      "37        1\n",
      "42        1\n",
      "46        1\n",
      "64        1\n",
      "63        1\n",
      "Name: number_emergency, dtype: int64\n",
      "-------------\n",
      "number_inpatient\n",
      "0     67630\n",
      "1     19521\n",
      "2      7566\n",
      "3      3411\n",
      "4      1622\n",
      "5       812\n",
      "6       480\n",
      "7       268\n",
      "8       151\n",
      "9       111\n",
      "10       61\n",
      "11       49\n",
      "12       34\n",
      "13       20\n",
      "14       10\n",
      "15        9\n",
      "16        6\n",
      "19        2\n",
      "17        1\n",
      "18        1\n",
      "21        1\n",
      "Name: number_inpatient, dtype: int64\n",
      "-------------\n",
      "diag_1\n",
      "428    6862\n",
      "414    6581\n",
      "786    4016\n",
      "410    3614\n",
      "486    3508\n",
      "       ... \n",
      "834       1\n",
      "833       1\n",
      "318       1\n",
      "477       1\n",
      "133       1\n",
      "Name: diag_1, Length: 717, dtype: int64\n",
      "-------------\n",
      "diag_2\n",
      "276    6752\n",
      "428    6662\n",
      "250    6071\n",
      "427    5036\n",
      "401    3736\n",
      "       ... \n",
      "894       1\n",
      "V25       1\n",
      "7         1\n",
      "703       1\n",
      "140       1\n",
      "Name: diag_2, Length: 749, dtype: int64\n",
      "-------------\n",
      "diag_3\n",
      "250    11555\n",
      "401     8289\n",
      "276     5175\n",
      "428     4577\n",
      "427     3955\n",
      "       ...  \n",
      "538        1\n",
      "877        1\n",
      "115        1\n",
      "872        1\n",
      "697        1\n",
      "Name: diag_3, Length: 790, dtype: int64\n",
      "-------------\n",
      "number_diagnoses\n",
      "9     49474\n",
      "5     11393\n",
      "8     10616\n",
      "7     10393\n",
      "6     10161\n",
      "4      5537\n",
      "3      2835\n",
      "2      1023\n",
      "1       219\n",
      "16       45\n",
      "10       17\n",
      "13       16\n",
      "11       11\n",
      "15       10\n",
      "12        9\n",
      "14        7\n",
      "Name: number_diagnoses, dtype: int64\n",
      "-------------\n",
      "max_glu_serum\n",
      "None    96420\n",
      "Norm     2597\n",
      ">200     1485\n",
      ">300     1264\n",
      "Name: max_glu_serum, dtype: int64\n",
      "-------------\n",
      "A1Cresult\n",
      "None    84748\n",
      ">8       8216\n",
      "Norm     4990\n",
      ">7       3812\n",
      "Name: A1Cresult, dtype: int64\n",
      "-------------\n",
      "metformin\n",
      "No        81778\n",
      "Steady    18346\n",
      "Up         1067\n",
      "Down        575\n",
      "Name: metformin, dtype: int64\n",
      "-------------\n",
      "repaglinide\n",
      "No        100227\n",
      "Steady      1384\n",
      "Up           110\n",
      "Down          45\n",
      "Name: repaglinide, dtype: int64\n",
      "-------------\n",
      "nateglinide\n",
      "No        101063\n",
      "Steady       668\n",
      "Up            24\n",
      "Down          11\n",
      "Name: nateglinide, dtype: int64\n",
      "-------------\n",
      "chlorpropamide\n",
      "No        101680\n",
      "Steady        79\n",
      "Up             6\n",
      "Down           1\n",
      "Name: chlorpropamide, dtype: int64\n",
      "-------------\n",
      "glimepiride\n",
      "No        96575\n",
      "Steady     4670\n",
      "Up          327\n",
      "Down        194\n",
      "Name: glimepiride, dtype: int64\n",
      "-------------\n",
      "acetohexamide\n",
      "No        101765\n",
      "Steady         1\n",
      "Name: acetohexamide, dtype: int64\n",
      "-------------\n",
      "glipizide\n",
      "No        89080\n",
      "Steady    11356\n",
      "Up          770\n",
      "Down        560\n",
      "Name: glipizide, dtype: int64\n",
      "-------------\n",
      "glyburide\n",
      "No        91116\n",
      "Steady     9274\n",
      "Up          812\n",
      "Down        564\n",
      "Name: glyburide, dtype: int64\n",
      "-------------\n",
      "tolbutamide\n",
      "No        101743\n",
      "Steady        23\n",
      "Name: tolbutamide, dtype: int64\n",
      "-------------\n",
      "pioglitazone\n",
      "No        94438\n",
      "Steady     6976\n",
      "Up          234\n",
      "Down        118\n",
      "Name: pioglitazone, dtype: int64\n",
      "-------------\n",
      "rosiglitazone\n",
      "No        95401\n",
      "Steady     6100\n",
      "Up          178\n",
      "Down         87\n",
      "Name: rosiglitazone, dtype: int64\n",
      "-------------\n",
      "acarbose\n",
      "No        101458\n",
      "Steady       295\n",
      "Up            10\n",
      "Down           3\n",
      "Name: acarbose, dtype: int64\n",
      "-------------\n",
      "miglitol\n",
      "No        101728\n",
      "Steady        31\n",
      "Down           5\n",
      "Up             2\n",
      "Name: miglitol, dtype: int64\n",
      "-------------\n",
      "troglitazone\n",
      "No        101763\n",
      "Steady         3\n",
      "Name: troglitazone, dtype: int64\n",
      "-------------\n",
      "tolazamide\n",
      "No        101727\n",
      "Steady        38\n",
      "Up             1\n",
      "Name: tolazamide, dtype: int64\n",
      "-------------\n",
      "examide\n",
      "No    101766\n",
      "Name: examide, dtype: int64\n",
      "-------------\n",
      "citoglipton\n",
      "No    101766\n",
      "Name: citoglipton, dtype: int64\n",
      "-------------\n",
      "insulin\n",
      "No        47383\n",
      "Steady    30849\n",
      "Down      12218\n",
      "Up        11316\n",
      "Name: insulin, dtype: int64\n",
      "-------------\n",
      "glyburide-metformin\n",
      "No        101060\n",
      "Steady       692\n",
      "Up             8\n",
      "Down           6\n",
      "Name: glyburide-metformin, dtype: int64\n",
      "-------------\n",
      "glipizide-metformin\n",
      "No        101753\n",
      "Steady        13\n",
      "Name: glipizide-metformin, dtype: int64\n",
      "-------------\n",
      "glimepiride-pioglitazone\n",
      "No        101765\n",
      "Steady         1\n",
      "Name: glimepiride-pioglitazone, dtype: int64\n",
      "-------------\n",
      "metformin-rosiglitazone\n",
      "No        101764\n",
      "Steady         2\n",
      "Name: metformin-rosiglitazone, dtype: int64\n",
      "-------------\n",
      "metformin-pioglitazone\n",
      "No        101765\n",
      "Steady         1\n",
      "Name: metformin-pioglitazone, dtype: int64\n",
      "-------------\n",
      "change\n",
      "No    54755\n",
      "Ch    47011\n",
      "Name: change, dtype: int64\n",
      "-------------\n",
      "diabetesMed\n",
      "Yes    78363\n",
      "No     23403\n",
      "Name: diabetesMed, dtype: int64\n",
      "-------------\n",
      "readmitted\n",
      "NO     54864\n",
      ">30    35545\n",
      "<30    11357\n",
      "Name: readmitted, dtype: int64\n",
      "-------------\n"
     ]
    }
   ],
   "source": [
    "for i in df.columns[2:]:\n",
    "    print(i)\n",
    "    print(df[i].value_counts())\n",
    "    print(\"-------------\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# After evaluating the data above we decided to perform the following dataengeniering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "\n",
    "1- Remove Encounter ID, Patient NUBR (unique ids, not needed)\n",
    "\n",
    "2- Race -> It is a categorical variable, therefore we will impute the missing valuees with the Mode.\n",
    "\n",
    "3- Age -> Bin in 3 categories, young, middle and old\n",
    "\n",
    "4- Weight -> Remove, too many unknowns, there is more than 85% of the data that is unknown.\n",
    "\n",
    "5- Admission ID Type -> Merging values 5, 8 into 6 (see mapping spreadsheet). To put all the missing values in one place.\n",
    "\n",
    "6- Admission Source -> Merging values 9, 15, 20, 21 into 17. To put all the missing values in one place.\n",
    "\n",
    "7- Discharge disposition -> remove observations where the discharge disposition is related to hospice or death since these will not add to the possibility of being readmitted \n",
    "\n",
    "8- DX -> Bing by using the ICD code in order to have the diagnosis of cardiovascular, respiratory, GI, Endocrinology, Diabetes, Neoplasis, Injury, Uninary and others\n",
    "\n",
    "9- Remove Payor Code - not helpful\n",
    "\n",
    "10- Remove Medical speciality - not helpful\n",
    "\n",
    "11- For the Categorical variables we will keep the none buckets, except examide and citoglipton since they only have one value\n",
    "\n",
    "12- Numerical features, some are normally distrbuted. We will use RobustScaler. The data that look non normaly distributed are the discrete numerical data, therefore we decide no to log transform the data,\n",
    "\n",
    "\n",
    "13- Readmitted -> Convert to binary as you will see in the code below."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# First drop the unnecesary columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First delate the columns that are id of the patients or encounter, so they are usefull for the model, and the columns of weight which is important but the amount of missing values is too much to make it an useful variable, there is 96% of missing values. We also delete the examide medication and citoglipton medication because there is only the no variable, there is not any yes value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "?            96.858479\n",
       "[75-100)      1.312816\n",
       "[50-75)       0.881434\n",
       "[100-125)     0.614154\n",
       "[125-150)     0.142484\n",
       "[25-50)       0.095317\n",
       "[0-25)        0.047167\n",
       "[150-175)     0.034393\n",
       "[175-200)     0.010809\n",
       ">200          0.002948\n",
       "Name: weight, dtype: float64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l=df[\"weight\"].value_counts(normalize=True)*100\n",
    "l"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['encounter_id', 'patient_nbr', 'race', 'gender', 'age', 'weight',\n",
       "       'admission_type_id', 'discharge_disposition_id', 'admission_source_id',\n",
       "       'time_in_hospital', 'payer_code', 'medical_specialty',\n",
       "       'num_lab_procedures', 'num_procedures', 'num_medications',\n",
       "       'number_outpatient', 'number_emergency', 'number_inpatient', 'diag_1',\n",
       "       'diag_2', 'diag_3', 'number_diagnoses', 'max_glu_serum', 'A1Cresult',\n",
       "       'metformin', 'repaglinide', 'nateglinide', 'chlorpropamide',\n",
       "       'glimepiride', 'acetohexamide', 'glipizide', 'glyburide', 'tolbutamide',\n",
       "       'pioglitazone', 'rosiglitazone', 'acarbose', 'miglitol', 'troglitazone',\n",
       "       'tolazamide', 'examide', 'citoglipton', 'insulin',\n",
       "       'glyburide-metformin', 'glipizide-metformin',\n",
       "       'glimepiride-pioglitazone', 'metformin-rosiglitazone',\n",
       "       'metformin-pioglitazone', 'change', 'diabetesMed', 'readmitted'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>race</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>admission_type_id</th>\n",
       "      <th>discharge_disposition_id</th>\n",
       "      <th>admission_source_id</th>\n",
       "      <th>time_in_hospital</th>\n",
       "      <th>num_lab_procedures</th>\n",
       "      <th>num_procedures</th>\n",
       "      <th>num_medications</th>\n",
       "      <th>...</th>\n",
       "      <th>tolazamide</th>\n",
       "      <th>insulin</th>\n",
       "      <th>glyburide-metformin</th>\n",
       "      <th>glipizide-metformin</th>\n",
       "      <th>glimepiride-pioglitazone</th>\n",
       "      <th>metformin-rosiglitazone</th>\n",
       "      <th>metformin-pioglitazone</th>\n",
       "      <th>change</th>\n",
       "      <th>diabetesMed</th>\n",
       "      <th>readmitted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Female</td>\n",
       "      <td>[0-10)</td>\n",
       "      <td>6</td>\n",
       "      <td>25</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>41</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Female</td>\n",
       "      <td>[10-20)</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>59</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>Up</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Ch</td>\n",
       "      <td>Yes</td>\n",
       "      <td>&gt;30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AfricanAmerican</td>\n",
       "      <td>Female</td>\n",
       "      <td>[20-30)</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>13</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>[30-40)</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>44</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>Up</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Ch</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>[40-50)</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>Steady</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Ch</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NO</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 43 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              race  gender      age  admission_type_id  \\\n",
       "0        Caucasian  Female   [0-10)                  6   \n",
       "1        Caucasian  Female  [10-20)                  1   \n",
       "2  AfricanAmerican  Female  [20-30)                  1   \n",
       "3        Caucasian    Male  [30-40)                  1   \n",
       "4        Caucasian    Male  [40-50)                  1   \n",
       "\n",
       "   discharge_disposition_id  admission_source_id  time_in_hospital  \\\n",
       "0                        25                    1                 1   \n",
       "1                         1                    7                 3   \n",
       "2                         1                    7                 2   \n",
       "3                         1                    7                 2   \n",
       "4                         1                    7                 1   \n",
       "\n",
       "   num_lab_procedures  num_procedures  num_medications  ...  tolazamide  \\\n",
       "0                  41               0                1  ...          No   \n",
       "1                  59               0               18  ...          No   \n",
       "2                  11               5               13  ...          No   \n",
       "3                  44               1               16  ...          No   \n",
       "4                  51               0                8  ...          No   \n",
       "\n",
       "   insulin  glyburide-metformin glipizide-metformin glimepiride-pioglitazone  \\\n",
       "0       No                   No                  No                       No   \n",
       "1       Up                   No                  No                       No   \n",
       "2       No                   No                  No                       No   \n",
       "3       Up                   No                  No                       No   \n",
       "4   Steady                   No                  No                       No   \n",
       "\n",
       "  metformin-rosiglitazone  metformin-pioglitazone change diabetesMed  \\\n",
       "0                      No                      No     No          No   \n",
       "1                      No                      No     Ch         Yes   \n",
       "2                      No                      No     No         Yes   \n",
       "3                      No                      No     Ch         Yes   \n",
       "4                      No                      No     Ch         Yes   \n",
       "\n",
       "  readmitted  \n",
       "0         NO  \n",
       "1        >30  \n",
       "2         NO  \n",
       "3         NO  \n",
       "4         NO  \n",
       "\n",
       "[5 rows x 43 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_reduced= df.drop(columns=['encounter_id','patient_nbr','weight','payer_code', 'medical_specialty', 'examide','citoglipton'])\n",
    "df_reduced.shape\n",
    "df_reduced.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Binning of the dig 1 diag 2 and dig 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to Categorize DX based on URL ...\n",
    "def bin_dx(code):\n",
    "    try:\n",
    "        code = float(code)\n",
    "    except:\n",
    "        code = 0\n",
    "\n",
    "    # Circulatory\n",
    "    if code in range(390,460) or code == 785:\n",
    "        return(\"Circulatory\")\n",
    "\n",
    "    # Respiratory\n",
    "    elif code in range(460,520) or code == 786:\n",
    "        return(\"Respiratory\")\n",
    "\n",
    "    # Digestive\n",
    "    elif code in range(520,580) or code == 787:\n",
    "        return(\"Digestive\")\n",
    "\n",
    "    # Diabetes\n",
    "    elif code >= 250 and code < 251:\n",
    "        return(\"Diabetes\")\n",
    "\n",
    "    # Injury\n",
    "    elif code in range(800,1000):\n",
    "        return(\"Injury\")\n",
    "\n",
    "    # Musculoskeletal\n",
    "    elif code in range(710,740):\n",
    "        return(\"Musculoskeletal\")\n",
    "\n",
    "    # Genitourinary\n",
    "    elif code in range(580,630) or code == 788:\n",
    "        return(\"Genitourinary\")\n",
    "\n",
    "    # Neoplasms\n",
    "    elif code in range(140,240):\n",
    "        return(\"Neoplasms\")\n",
    "    elif code in range(780,783) or code == 784:\n",
    "        return(\"Neoplasms\")\n",
    "    elif code in range(790,800):\n",
    "        return(\"Neoplasms\")\n",
    "    elif code in range(240,250):\n",
    "        return(\"Neoplasms\")\n",
    "    elif code in range(251,280):\n",
    "        return(\"Neoplasms\")\n",
    "    elif code in range(680,710):\n",
    "        return(\"Neoplasms\")\n",
    "    elif code in range(1,140):\n",
    "        return(\"Neoplasms\")\n",
    "    elif code in range(290,320):\n",
    "        return(\"Neoplasms\")\n",
    "\n",
    "    # Other\n",
    "    else:\n",
    "        return(\"Other\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Circulatory        30437\n",
      "Neoplasms          16318\n",
      "Respiratory        14423\n",
      "Digestive           9475\n",
      "Diabetes            8757\n",
      "Injury              6974\n",
      "Other               5308\n",
      "Genitourinary       5117\n",
      "Musculoskeletal     4957\n",
      "Name: Diag1, dtype: int64\n",
      "\n",
      "Circulatory        31881\n",
      "Neoplasms          21462\n",
      "Diabetes           12794\n",
      "Respiratory        10895\n",
      "Genitourinary       8376\n",
      "Other               7996\n",
      "Digestive           4170\n",
      "Injury              2428\n",
      "Musculoskeletal     1764\n",
      "Name: Diag2, dtype: int64\n",
      "\n",
      "Circulatory        30306\n",
      "Neoplasms          20985\n",
      "Diabetes           17157\n",
      "Other              11489\n",
      "Respiratory         7358\n",
      "Genitourinary       6680\n",
      "Digestive           3930\n",
      "Injury              1946\n",
      "Musculoskeletal     1915\n",
      "Name: Diag3, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "df_reduced[\"Diag1\"] = df_reduced[\"diag_1\"].apply(bin_dx)\n",
    "df_reduced[\"Diag2\"] = df_reduced[\"diag_2\"].apply(bin_dx)\n",
    "df_reduced[\"Diag3\"] = df_reduced[\"diag_3\"].apply(bin_dx)\n",
    "\n",
    "\n",
    "print(df_reduced[\"Diag1\"].value_counts())\n",
    "print(\"\")\n",
    "print(df_reduced[\"Diag2\"].value_counts())\n",
    "print(\"\")\n",
    "print(df_reduced[\"Diag3\"].value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_reduced = df_reduced.drop(columns=['diag_1', 'diag_2', 'diag_3'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# We will bin the age variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[70-80)     26068\n",
      "[60-70)     22483\n",
      "[50-60)     17256\n",
      "[80-90)     17197\n",
      "[40-50)      9685\n",
      "[30-40)      3775\n",
      "[90-100)     2793\n",
      "[20-30)      1657\n",
      "[10-20)       691\n",
      "[0-10)        161\n",
      "Name: age, dtype: int64\n",
      "Old       68541\n",
      "Midage    26941\n",
      "Young      6284\n",
      "Name: age, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(df_reduced[\"age\"].value_counts())\n",
    "di = {'[70-80)': 'Old', '[60-70)': 'Old', '[80-90)': 'Old', '[90-100)': 'Old', '[50-60)': 'Midage', '[40-50)': 'Midage', '[30-40)': 'Young', '[20-30)': 'Young', '[10-20)': 'Young', '[0-10)': 'Young'}\n",
    "\n",
    "df_reduced[\"age\"].replace(di, inplace=True)\n",
    "print(df_reduced[\"age\"].value_counts())\n",
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# We will replace the missing values from the reminder column"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I there is a discrete or categorical values we will replace with the mode, if there is a continous variable we will replace with the mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reagenge the id coluns can be useful in a way that all the missing values is place in one place and make it categorical\n",
    "# admission_type_id\n",
    "df_reduced.admission_type_id.replace(to_replace = 5, value = 6, inplace = True)\n",
    "df_reduced.admission_type_id.replace(to_replace = 8, value = 6, inplace = True)\n",
    "df_reduced.admission_type_id = df_reduced.admission_type_id.astype(str)\n",
    "\n",
    "# admission_source_id\n",
    "df_reduced.admission_source_id.replace(to_replace = 9, value = 17, inplace = True)\n",
    "df_reduced.admission_source_id.replace(to_replace = 15, value = 17, inplace = True)\n",
    "df_reduced.admission_source_id.replace(to_replace = 20, value = 17, inplace = True)\n",
    "df_reduced.admission_source_id.replace(to_replace = 21, value = 17, inplace = True)\n",
    "df_reduced.admission_source_id= df_reduced.admission_source_id.astype(str)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Next, we remove observations where the discharge disposition is related to hospice or death since these will not add to the possibility of being readmitted.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1     60234\n",
      "3     13954\n",
      "6     12902\n",
      "18     3691\n",
      "2      2128\n",
      "22     1993\n",
      "5      1184\n",
      "25      989\n",
      "4       815\n",
      "7       623\n",
      "23      412\n",
      "28      139\n",
      "8       108\n",
      "15       63\n",
      "24       48\n",
      "9        21\n",
      "17       14\n",
      "16       11\n",
      "10        6\n",
      "27        5\n",
      "12        3\n",
      "Name: discharge_disposition_id, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "indexNames = df_reduced[df_reduced['discharge_disposition_id'].isin([11,13,14,19,20,21])].index\n",
    "df_reduced.drop(indexNames, inplace=True)\n",
    "\n",
    "print(df_reduced['discharge_disposition_id'].value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Make the result binary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0    88029\n",
      "1    11314\n",
      "Name: readmitted, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "di = {'NO': 0, '>30': 0, '<30': 1}\n",
    "\n",
    "df_reduced[\"readmitted\"].replace(di, inplace=True)\n",
    "print(df_reduced[\"readmitted\"].value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot the readmission variable to show a big inbalance in the response"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df_reduced[\"readmitted\"])\n",
    "my_xtick = [0, 0.5, 1]\n",
    "my_xtick_labels = ['0' , '', '1']\n",
    "plt.xticks(my_xtick, my_xtick_labels, fontsize=8)\n",
    "plt.xlabel(\"Readmission\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# One Hot encode the categorical values and scale the continous variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"categorical_values= ['race', 'gender', 'max_glu_serum', 'A1Cresult', 'metformin', 'repaglinide', 'nateglinide','chlorpropamide',\\n                     'glimepiride', 'acetohexamide', 'glipizide','glyburide', 'tolbutamide', 'pioglitazone', 'rosiglitazone', 'acarbose','miglitol', 'troglitazone', 'tolazamide', 'insulin',\\n                    'glyburide-metformin', 'glipizide-metformin','glimepiride-pioglitazone', 'metformin-rosiglitazone',\\n                    'metformin-pioglitazone', 'change', 'diabetesMed', 'Diag1', 'Diag2','Diag3', 'Age', 'admission_type_id', 'discharge_disposition_id','admission_source_id']\\n\\n\\n\\nnumeric_value= ['time_in_hospital', 'num_lab_procedures','num_procedures', 'num_medications', 'number_outpatient',\\n                'number_emergency', 'number_inpatient', 'number_diagnoses',]\\n\""
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"categorical_values= ['race', 'gender', 'max_glu_serum', 'A1Cresult', 'metformin', 'repaglinide', 'nateglinide','chlorpropamide',\n",
    "                     'glimepiride', 'acetohexamide', 'glipizide','glyburide', 'tolbutamide', 'pioglitazone', 'rosiglitazone', 'acarbose','miglitol', 'troglitazone', 'tolazamide', 'insulin',\n",
    "                    'glyburide-metformin', 'glipizide-metformin','glimepiride-pioglitazone', 'metformin-rosiglitazone',\n",
    "                    'metformin-pioglitazone', 'change', 'diabetesMed', 'Diag1', 'Diag2','Diag3', 'Age', 'admission_type_id', 'discharge_disposition_id','admission_source_id']\n",
    "\n",
    "\n",
    "\n",
    "numeric_value= ['time_in_hospital', 'num_lab_procedures','num_procedures', 'num_medications', 'number_outpatient',\n",
    "                'number_emergency', 'number_inpatient', 'number_diagnoses',]\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"from sklearn.preprocessing import OneHotEncoder\\n# categorical\\nohe = OneHotEncoder()\\nX_object = df_reduced.select_dtypes('object')\\nohe.fit(X_object)\\ncodes = ohe.transform(X_object).toarray()\\nfeature_names = ohe.get_feature_names(X_object.columns)\\n\\n# integer scaled\\nx_int= df_reduced.select_dtypes(exclude='object')\\nmin_max_scaler = RobustScaler()\\nx_int = min_max_scaler.fit_transform(x_int)\\nfeature_names_i =df_reduced.select_dtypes(exclude='object').columns \\n\\nfr = pd.concat([pd.DataFrame(x_int ,columns=feature_names_i), \\n               pd.DataFrame(codes,columns=feature_names).astype(int)], axis=1)\""
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"from sklearn.preprocessing import OneHotEncoder\n",
    "# categorical\n",
    "ohe = OneHotEncoder()\n",
    "X_object = df_reduced.select_dtypes('object')\n",
    "ohe.fit(X_object)\n",
    "codes = ohe.transform(X_object).toarray()\n",
    "feature_names = ohe.get_feature_names(X_object.columns)\n",
    "\n",
    "# integer scaled\n",
    "x_int= df_reduced.select_dtypes(exclude='object')\n",
    "min_max_scaler = RobustScaler()\n",
    "x_int = min_max_scaler.fit_transform(x_int)\n",
    "feature_names_i =df_reduced.select_dtypes(exclude='object').columns \n",
    "\n",
    "fr = pd.concat([pd.DataFrame(x_int ,columns=feature_names_i), \n",
    "               pd.DataFrame(codes,columns=feature_names).astype(int)], axis=1)\"\"\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Bulding the model with the pipeline using one hot encoder inside the pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(99343, 43)\n",
      "(99343, 42)\n",
      "(99343,)\n"
     ]
    }
   ],
   "source": [
    "categorical_features = ['race', 'gender', 'max_glu_serum', 'A1Cresult', 'metformin', 'repaglinide', 'nateglinide','chlorpropamide',\n",
    "                        'glimepiride', 'acetohexamide', 'glipizide','glyburide', 'tolbutamide', 'pioglitazone', 'rosiglitazone', \n",
    "                        'acarbose','miglitol', 'troglitazone', 'tolazamide', 'insulin', 'glyburide-metformin', 'glipizide-metformin',\n",
    "                        'glimepiride-pioglitazone', 'metformin-rosiglitazone', 'metformin-pioglitazone', 'change', 'diabetesMed',\n",
    "                        'Diag1', 'Diag2','Diag3', 'age', 'admission_type_id', 'discharge_disposition_id','admission_source_id']\n",
    "\n",
    "numeric_features = ['time_in_hospital', 'num_lab_procedures','num_procedures', 'num_medications', 'number_outpatient', \n",
    "                    'number_emergency', 'number_inpatient', 'number_diagnoses']\n",
    "\n",
    "X = df_reduced[categorical_features + numeric_features]\n",
    "y = df_reduced['readmitted']\n",
    "\n",
    "# Verify shapes\n",
    "print(df_reduced.shape)\n",
    "print(X.shape)\n",
    "print(y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/fabiosavorgnan/opt/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross-validation score: 0.6446633628806032\n",
      "Test score: 0.6508134127937846\n"
     ]
    }
   ],
   "source": [
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,\n",
    "                                                    y,\n",
    "                                                    test_size=0.2,\n",
    "                                                    stratify=y,\n",
    "                                                    random_state=11)\n",
    "\n",
    "numeric_transformer = Pipeline(steps=[\n",
    "    ('imputer', SimpleImputer(missing_values=np.nan, strategy=\"median\")),\n",
    "    ('scaler', RobustScaler(with_centering=False))])\n",
    "\n",
    "categorical_transformer = Pipeline(steps=[\n",
    "    ('imputer', SimpleImputer(missing_values=np.nan, strategy=\"most_frequent\")),\n",
    "    ('onehot', OneHotEncoder(handle_unknown='ignore'))])\n",
    "\n",
    "preprocessor = ColumnTransformer(\n",
    "    transformers=[\n",
    "        ('num', numeric_transformer, numeric_features),\n",
    "        ('cat', categorical_transformer, categorical_features)])\n",
    "\n",
    "\n",
    "pipeline = imbpipeline(steps = [['preprocessor', preprocessor],\n",
    "                                ['smote', SMOTE(random_state=11)],\n",
    "                                ##['scaler', RobustScaler(with_centering=False)],\n",
    "                                ['classifier', LogisticRegression(random_state=11,\n",
    "                                                                  max_iter=1000, solver= \"saga\")]])\n",
    "\n",
    "stratified_kfold = StratifiedKFold(n_splits=3,\n",
    "                                       shuffle=True,\n",
    "                                       random_state=11)\n",
    "\n",
    "param_grid = {'classifier__C': [0.001, 0.01, 0.1, 1, 10, 100, 1000],\n",
    "              'classifier__penalty': ['l1', 'l2']}\n",
    "\n",
    "scorers = {\n",
    "    'precision_score': make_scorer(precision_score),\n",
    "    'recall_score': make_scorer(recall_score),\n",
    "    'accuracy_score': make_scorer(accuracy_score)\n",
    "}\n",
    "              \n",
    "grid_search = GridSearchCV(estimator=pipeline,\n",
    "                           param_grid=param_grid,\n",
    "                           scoring='roc_auc',\n",
    "                           #scoring=scorers,\n",
    "                           cv=stratified_kfold,\n",
    "                           n_jobs=-1)\n",
    "\n",
    "grid_search.fit(X_train, y_train)\n",
    "cv_score = grid_search.best_score_\n",
    "test_score = grid_search.score(X_test, y_test)\n",
    "print(f'Cross-validation score: {cv_score}\\nTest score: {test_score}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>classifier__C</th>\n",
       "      <th>classifier__penalty</th>\n",
       "      <th>test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.001</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.644475</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.001</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.643333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.010</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.644663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.010</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.640416</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.100</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.639719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.100</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.638522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.000</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.637907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1.000</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.637669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>10.000</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.637558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10.000</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.637535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>100.000</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.637523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>100.000</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.637521</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1000.000</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.637520</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1000.000</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.637519</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    classifier__C classifier__penalty  test_score\n",
       "0           0.001                  l1    0.644475\n",
       "1           0.001                  l2    0.643333\n",
       "2           0.010                  l1    0.644663\n",
       "3           0.010                  l2    0.640416\n",
       "4           0.100                  l1    0.639719\n",
       "5           0.100                  l2    0.638522\n",
       "6           1.000                  l1    0.637907\n",
       "7           1.000                  l2    0.637669\n",
       "8          10.000                  l1    0.637558\n",
       "9          10.000                  l2    0.637535\n",
       "10        100.000                  l1    0.637523\n",
       "11        100.000                  l2    0.637521\n",
       "12       1000.000                  l1    0.637520\n",
       "13       1000.000                  l2    0.637519"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_LR = pd.DataFrame(grid_search.cv_results_['params'])\n",
    "results_LR['test_score'] = grid_search.cv_results_['mean_test_score']\n",
    "results_LR "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Print the grid best stimator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'classifier__C': 0.01, 'classifier__penalty': 'l1'}\n"
     ]
    }
   ],
   "source": [
    "print(grid_search.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using the grid best stimator to predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_hat= grid_search.best_estimator_.predict(X_test)\n",
    "y_hat"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Best classifier coef"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"classifier = ovr.named_steps['classifier']\\nclassifier.coef_.shape\""
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"classifier = ovr.named_steps['classifier']\n",
    "classifier.coef_.shape\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/fabiosavorgnan/opt/anaconda3/lib/python3.7/site-packages/dask/dataframe/utils.py:14: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
      "  import pandas.util.testing as tm\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "    <style>\n",
       "    table.eli5-weights tr:hover {\n",
       "        filter: brightness(85%);\n",
       "    }\n",
       "</style>\n",
       "\n",
       "\n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "        \n",
       "\n",
       "    \n",
       "\n",
       "        \n",
       "            \n",
       "                \n",
       "                \n",
       "    \n",
       "        <p style=\"margin-bottom: 0.5em; margin-top: 0em\">\n",
       "            <b>\n",
       "    \n",
       "        y=1\n",
       "    \n",
       "</b>\n",
       "\n",
       "top features\n",
       "        </p>\n",
       "    \n",
       "    <table class=\"eli5-weights\"\n",
       "           style=\"border-collapse: collapse; border: none; margin-top: 0em; table-layout: auto; margin-bottom: 2em;\">\n",
       "        <thead>\n",
       "        <tr style=\"border: none;\">\n",
       "            \n",
       "                <th style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\" title=\"Feature weights. Note that weights do not account for feature value scales, so if feature values have different scales, features with highest weights might not be the most important.\">\n",
       "                    Weight<sup>?</sup>\n",
       "                </th>\n",
       "            \n",
       "            <th style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">Feature</th>\n",
       "            \n",
       "        </tr>\n",
       "        </thead>\n",
       "        <tbody>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 80.00%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +1.042\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        discharge_disposition_id_22\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 90.79%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.344\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        A1Cresult_None\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 91.74%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.295\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        race_Caucasian\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 91.92%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.285\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        discharge_disposition_id_5\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 92.10%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.277\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        number_inpatient\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 94.08%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.183\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag2_Neoplasms\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 94.30%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.173\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        admission_source_id_1\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 94.36%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.171\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        admission_source_id_7\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 94.62%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.160\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag1_Circulatory\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 94.91%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.148\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag2_Circulatory\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 95.00%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.144\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        number_diagnoses\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 95.15%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.138\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        race_AfricanAmerican\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 96.03%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.104\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag2_Diabetes\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 96.11%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.100\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        admission_type_id_1\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 96.67%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.080\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        insulin_Down\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 97.19%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.063\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        num_lab_procedures\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 97.35%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.058\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        discharge_disposition_id_3\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 97.91%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.042\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        num_medications\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 98.17%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        +0.034\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag3_Circulatory\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "        \n",
       "            <tr style=\"background-color: hsl(120, 100.00%, 98.17%); border: none;\">\n",
       "                <td colspan=\"2\" style=\"padding: 0 0.5em 0 0.5em; text-align: center; border: none; white-space: nowrap;\">\n",
       "                    <i>&hellip; 6 more positive &hellip;</i>\n",
       "                </td>\n",
       "            </tr>\n",
       "        \n",
       "\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 98.24%); border: none;\">\n",
       "                <td colspan=\"2\" style=\"padding: 0 0.5em 0 0.5em; text-align: center; border: none; white-space: nowrap;\">\n",
       "                    <i>&hellip; 8 more negative &hellip;</i>\n",
       "                </td>\n",
       "            </tr>\n",
       "        \n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 98.24%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.032\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        gender_Female\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 97.95%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.040\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag2_Respiratory\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 97.94%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.040\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag2_Other\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 97.84%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.043\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        number_outpatient\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 97.67%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.048\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        admission_type_id_6\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 97.63%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.049\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        metformin-rosiglitazone_No\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 97.61%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.050\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag1_Musculoskeletal\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 97.25%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.061\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        A1Cresult_Norm\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 96.70%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.079\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        num_procedures\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 96.40%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.090\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        max_glu_serum_None\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 96.37%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.091\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        diabetesMed_No\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 96.23%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.096\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        metformin_Steady\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 96.22%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.097\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag3_Other\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 96.06%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.102\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag3_Musculoskeletal\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 95.83%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.111\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        change_Ch\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 95.50%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.124\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        discharge_disposition_id_23\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 95.48%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.124\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        Diag1_Respiratory\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 95.19%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.136\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        insulin_Steady\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 95.14%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.138\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        acetohexamide_No\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 94.70%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.156\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        discharge_disposition_id_6\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 94.53%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.164\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        rosiglitazone_Steady\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 94.26%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.175\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        age_Young\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 94.04%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.185\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        glimepiride-pioglitazone_No\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 94.00%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.186\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        metformin_Up\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 93.01%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.232\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        age_Midage\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 92.92%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.237\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        glimepiride_Steady\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 92.91%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.237\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        pioglitazone_Steady\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 92.78%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.243\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        metformin-pioglitazone_No\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 92.57%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.254\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        discharge_disposition_id_1\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 91.81%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.291\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        insulin_No\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "            <tr style=\"background-color: hsl(0, 100.00%, 90.11%); border: none;\">\n",
       "    <td style=\"padding: 0 1em 0 0.5em; text-align: right; border: none;\">\n",
       "        -0.381\n",
       "    </td>\n",
       "    <td style=\"padding: 0 0.5em 0 0.5em; text-align: left; border: none;\">\n",
       "        &lt;BIAS&gt;\n",
       "    </td>\n",
       "    \n",
       "</tr>\n",
       "        \n",
       "\n",
       "        </tbody>\n",
       "    </table>\n",
       "\n",
       "            \n",
       "        \n",
       "\n",
       "        \n",
       "\n",
       "\n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "    \n",
       "\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "Explanation(estimator=\"LogisticRegression(C=0.01, max_iter=1000, penalty='l1', random_state=11,\\n                   solver='saga')\", description=\"\\nFeatures with largest coefficients.\\nCaveats:\\n1. Be careful with features which are not\\n   independent - weights don't show their importance.\\n2. If scale of input features is different then scale of coefficients\\n   will also be different, making direct comparison between coefficient values\\n   incorrect.\\n3. Depending on regularization, rare features sometimes may have high\\n   coefficients; this doesn't mean they contribute much to the\\n   classification result for most examples.\\n\", error=None, method='linear model', is_regression=False, targets=[TargetExplanation(target=1, feature_weights=FeatureWeights(pos=[FeatureWeight(feature='discharge_disposition_id_22', weight=1.0424752705003992, std=None, value=None), FeatureWeight(feature='A1Cresult_None', weight=0.3443882193442531, std=None, value=None), FeatureWeight(feature='race_Caucasian', weight=0.29466385567181635, std=None, value=None), FeatureWeight(feature='discharge_disposition_id_5', weight=0.28537403005979306, std=None, value=None), FeatureWeight(feature='number_inpatient', weight=0.2766188837337321, std=None, value=None), FeatureWeight(feature='Diag2_Neoplasms', weight=0.18291420725438012, std=None, value=None), FeatureWeight(feature='admission_source_id_1', weight=0.17347437482101835, std=None, value=None), FeatureWeight(feature='admission_source_id_7', weight=0.1710176344400644, std=None, value=None), FeatureWeight(feature='Diag1_Circulatory', weight=0.1598725832448925, std=None, value=None), FeatureWeight(feature='Diag2_Circulatory', weight=0.1475715807249754, std=None, value=None), FeatureWeight(feature='number_diagnoses', weight=0.14380498895222274, std=None, value=None), FeatureWeight(feature='race_AfricanAmerican', weight=0.13769916412223032, std=None, value=None), FeatureWeight(feature='Diag2_Diabetes', weight=0.1035700523153609, std=None, value=None), FeatureWeight(feature='admission_type_id_1', weight=0.10037404177400949, std=None, value=None), FeatureWeight(feature='insulin_Down', weight=0.08038035487351794, std=None, value=None), FeatureWeight(feature='num_lab_procedures', weight=0.0630291307981887, std=None, value=None), FeatureWeight(feature='discharge_disposition_id_3', weight=0.05799471375324819, std=None, value=None), FeatureWeight(feature='num_medications', weight=0.04150269460585795, std=None, value=None), FeatureWeight(feature='Diag3_Circulatory', weight=0.03433830120144409, std=None, value=None)], neg=[FeatureWeight(feature='<BIAS>', weight=-0.38147410881473537, std=None, value=None), FeatureWeight(feature='insulin_No', weight=-0.2910950549136945, std=None, value=None), FeatureWeight(feature='discharge_disposition_id_1', weight=-0.2535538993270248, std=None, value=None), FeatureWeight(feature='metformin-pioglitazone_No', weight=-0.24332286312482349, std=None, value=None), FeatureWeight(feature='pioglitazone_Steady', weight=-0.2370979194788936, std=None, value=None), FeatureWeight(feature='glimepiride_Steady', weight=-0.23664357347797377, std=None, value=None), FeatureWeight(feature='age_Midage', weight=-0.23231297392833308, std=None, value=None), FeatureWeight(feature='metformin_Up', weight=-0.18646631377319225, std=None, value=None), FeatureWeight(feature='glimepiride-pioglitazone_No', weight=-0.18498624937816432, std=None, value=None), FeatureWeight(feature='age_Young', weight=-0.17505642131744792, std=None, value=None), FeatureWeight(feature='rosiglitazone_Steady', weight=-0.16367949701186787, std=None, value=None), FeatureWeight(feature='discharge_disposition_id_6', weight=-0.15628228572725725, std=None, value=None), FeatureWeight(feature='acetohexamide_No', weight=-0.13803980682771538, std=None, value=None), FeatureWeight(feature='insulin_Steady', weight=-0.13604899053381284, std=None, value=None), FeatureWeight(feature='Diag1_Respiratory', weight=-0.12447400126751909, std=None, value=None), FeatureWeight(feature='discharge_disposition_id_23', weight=-0.12385505536502861, std=None, value=None), FeatureWeight(feature='change_Ch', weight=-0.11093693950892627, std=None, value=None), FeatureWeight(feature='Diag3_Musculoskeletal', weight=-0.10243855291307831, std=None, value=None), FeatureWeight(feature='Diag3_Other', weight=-0.09653387234612282, std=None, value=None), FeatureWeight(feature='metformin_Steady', weight=-0.09594507250116902, std=None, value=None), FeatureWeight(feature='diabetesMed_No', weight=-0.09096213838647779, std=None, value=None), FeatureWeight(feature='max_glu_serum_None', weight=-0.08986052958897764, std=None, value=None), FeatureWeight(feature='num_procedures', weight=-0.07948375284994041, std=None, value=None), FeatureWeight(feature='A1Cresult_Norm', weight=-0.06110024674682506, std=None, value=None), FeatureWeight(feature='Diag1_Musculoskeletal', weight=-0.050190467641076396, std=None, value=None), FeatureWeight(feature='metformin-rosiglitazone_No', weight=-0.04945421110715306, std=None, value=None), FeatureWeight(feature='admission_type_id_6', weight=-0.04826756417873535, std=None, value=None), FeatureWeight(feature='number_outpatient', weight=-0.04348493460107751, std=None, value=None), FeatureWeight(feature='Diag2_Other', weight=-0.040457219599453864, std=None, value=None), FeatureWeight(feature='Diag2_Respiratory', weight=-0.04016288952350226, std=None, value=None), FeatureWeight(feature='gender_Female', weight=-0.03243293691092763, std=None, value=None)], pos_remaining=6, neg_remaining=8), proba=None, score=None, weighted_spans=None, heatmap=None)], feature_importances=None, decision_tree=None, highlight_spaces=None, transition_features=None, image=None)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import eli5\n",
    "\n",
    "onehot_columns = list(grid_search.best_estimator_.named_steps['preprocessor'].named_transformers_['cat'].named_steps['onehot'].get_feature_names(input_features=categorical_features))\n",
    "numeric_features_list = list(numeric_features)\n",
    "numeric_features_list.extend(onehot_columns)\n",
    "\n",
    "eli5.explain_weights(grid_search.best_estimator_.named_steps['classifier'], top=50, feature_names=numeric_features_list)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Performance of the modeel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Misclassified samples: 6651\n",
      "Accuracy: 0.67\n",
      "Precision= 0.6652574362071568\n",
      "Recal= 0.6652574362071568\n",
      "fscore= 0.7243598423040026\n",
      "Confusion Matrix = [[12000  5606]\n",
      " [ 1045  1218]]\n"
     ]
    }
   ],
   "source": [
    "# how did our model perform?\n",
    "count_misclassified = (y_test != y_hat).sum()\n",
    "print('Misclassified samples: {}'.format(count_misclassified))\n",
    "accuracy = mt.accuracy_score(y_test, y_hat)\n",
    "print('Accuracy: {:.2f}'.format(accuracy))\n",
    "precision = mt.precision_score(y_test, y_hat, average='micro')\n",
    "print(\"Precision=\", precision)\n",
    "recal = mt.recall_score(y_test, y_hat, average='micro')\n",
    "print(\"Recal=\", recal)\n",
    "fscore = mt.f1_score(y_test, y_hat, average='weighted')\n",
    "print(\"fscore=\",fscore )\n",
    "# Confusion matrix\n",
    "cm= confusion_matrix(y_test, y_hat)\n",
    "print(\"Confusion Matrix =\", cm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[12000  5606]\n",
      " [ 1045  1218]]\n",
      "Normalized confusion matrix\n",
      "[[0.68 0.32]\n",
      " [0.46 0.54]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATwAAAEWCAYAAAD7MitWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/d3fzzAAAACXBIWXMAAAsTAAALEwEAmpwYAAAlBUlEQVR4nO3deZwV1Zn/8c+3u0EREFBEXNgkiLtGEYNEgzE66Ji4JVGTiXFMRklinEw2zcxkkjGT1eQXZ6IZooYxxgWNW1CJaFSiRiOIOwiRIEsLCo2Agig2PL8/qhoul15uwe3ue6nvm9d9Ucupqufe2/30OXWq6igiMDPLg5rODsDMrKM44ZlZbjjhmVluOOGZWW444ZlZbjjhmVluOOFVCElTJX0+nf60pPvLvP/BkkJSXTn328YxJen/JK2QNG0b9nOMpDnljK2zSBooabWk2s6OJY9yk/AkzZf0uqTuBcs+L2lqJ4bVrIi4MSJO7Ow4yuCDwAnA3hExcmt3EhGPRsTw8oXVPtKfsY+0ViYiFkZEj4hY31Fx2Sa5SXipOuCft3Unac0lb5/d1hgEzI+INZ0dSCXoyNq1NS9vv7SXA1+X1Lu5lZKOljRd0qr0/6ML1k2V9H1JfwbeBvZJm4hflPSypLckfU/SUElPSHpT0q2Suqbb95F0j6RlaRPvHkl7txDHeZIeS6e/mTaBml7vSbouXddL0q8lLZH0qqT/amoqSaqV9FNJDZLmAX/f2gcjaYCkO9L4lku6Ml1eI+nfJS2QtFTS9ZJ6peuamsmflbQwPda/pes+B1wLjErj/s/C91Vw3JD0vnT6ZEmz0s/yVUlfT5ePkVRfsM3+6fexUtJMSR8rWHedpKsk3Zvu50lJQ1t4z03x/6OkRen3Mk7SkZKeT/d/ZUH5oZIeSj+fBkk3Nv0sSfotMBC4O32/3yzY/+ckLQQeKlhWJ2kXSfWSPpruo4ekuZLObe27sm0QEbl4AfOBjwB3AP+VLvs8MDWd3gVYAXyGpCZ4Tjq/a7p+KrAQODBd3wUIYBKwc7r8XeBBYB+gFzAL+Gy6/a7AmcBOQE/gd8BdBfFNBT6fTp8HPNbMexgALAZOTufvAn4FdAf6AdOAC9N144DZ6Ta7AA+n8dY1s99a4Dng5+m+dgQ+mK47H5ibvqce6ef323Td4HSf1wDdgEPTz2D/5t5Hc+8r3f596fQS4Jh0ug9weDo9BqhPp7uk8fwr0BX4MPAWMDxdfx3wBjAy/Z5uBCa28DPRFP/49D2fCLyTfq79gL2ApcCH0vLvI2mi7wDsBjwCXFH8M9bM/q9PP9duBcvq0jInAq+lx7sGuK2zf1e251enB9Bhb3RTwjsIWJX+wBYmvM8A04q2eQI4L52eClxWtD6A0QXzM4BLCuZ/VvgLUbTtYcCKgvmptJLw0l+WjfsHdk+TS7eCMucAD6fTDwHjCtadSMsJbxSwrIV1DwJfLJgfDryXJpOmX969C9ZPA85u7n208L4KE95C4EJg56IyY9iU8I5JE0RNwfqbge+m09cB1xasOxmY3cJ30BT/XgXLlgNnFczfDnylhe1PA54p/hlrZv/7NLOsrmDZL4AXSP6Y7drZvyvb8ytvTVoi4kXgHuDSolV7AguKli0g+SvfZFEzu3y9YHptM/M9ACTtJOlXadPwTZLaQW+V3lv3a2BORPw4nR9EUttZkja9VpLU9voVvJ/CeIvfW6EBwIKIaGxmXfHnsoAk2e1esOy1gum3Sd/zVjiTJEEtkPQnSaNaiGdRRGwoiqnwe8oaT6nfYT9JE9Pm9pvADUDfNvYNzf/cFLqa5A/x/0XE8hL2Z1spdwkv9R3gn9j8l2QxSRIpNBB4tWB+Wx4t8zWS2tFREbEzcGy6XG1tKOnSdNvPFSxeRFLD6xsRvdPXzhFxYLp+CUkiazKwlUMsAgaq+ZPqxZ/LQKCRzZNCqdaQNOkBkNS/cGVETI+IU0mS9l3ArS3EM0CbdxoVf0/t5YckPwOHpN/hP7D599fSz0eLPzfpH7xfkTR7v9B0PtPaRy4TXkTMBW4BLi5YPBnYV9Kn0hPKZwEHkNQGy6EnSW1hpaRdSJJumySdlMZ5WkSsLXgPS4D7gZ9J2jntXBgq6UNpkVuBiyXtLakPW9ZoC00jSZA/ktRd0o6SRqfrbgb+RdIQST2AHwC3tFAbbMtzwIGSDpO0I/DdgvfZVcn1h70i4j3gTaC5SzeeJEmc35TURdIY4KPAxK2IJ6uewGqS73Av4BtF618nOdeZxb+m/58P/BS4PkOt3zLKZcJLXUZyIhmAtClxCklNbDnwTeCUiGgo0/GuIDkP1wD8BbivxO3OIjnf+JI29dSOT9edS3LifhZJB8ttwB7pumuAKSRJ5mmSzoZmRXJN2EdJTsovBOrT4wJMAH5L0gR/heSk/pdLjL34OH8l+dz/CLwMPFZU5DPA/LS5OI6kBlW8j3XAx4CTSD7LXwLnRsTsrYkpo/8EDic5B3wvW36mPwT+PT3F8PW2dibpCOCrJPGvB35MUhts7Y+TbQOlJ03NzLZ7ea7hmVnOOOGZWW444ZlZbjjhmVluVNTNzKrrFuras7PDsAzev39rl/dZpVmwYD4NDQ1tXvvZmtqdB0U0rm27IBBrl02JiLHbcrxyqqyE17UnOwz/ZGeHYRn8+ckr2y5kFWP0USO2eR/RuLbk39N3nr2qlDtROkxFJTwzqwaCKn06mhOemWUjoKY6bwZxwjOz7LRNpwE7jROemWXkJq2Z5YlreGaWC8I1PDPLC7mGZ2Y54l5aM8sHd1qYWV4IN2nNLEdcwzOzfHCT1szyQkCtOy3MLC+q9BxeddZLzawTpU3aUl5t7UkaK2mOpLnp+MvNlRkj6VlJMyX9qWD5fEkvpOueKiVy1/DMLLsy1PDS8XevAk4gGRp0uqRJETGroExvkqE4x0bEQkn9inZzXJahVF3DM7PsylPDGwnMjYh56XjDE4FTi8p8CrgjIhYCRMTSbQnbCc/MspFKf0FfSU8VvC4o2NNewKKC+fp0WaF9gT6SpkqaIencgnUB3J8uv4ASuElrZtmVfmtZQ0S09Fz55trFUTRfBxwBHA90A56Q9JeI+CswOiIWp83cByTNjohHWg271KjNzBJl67SoBwYUzO8NLG6mzH0RsSY9V/cIcChARCxO/18K3EnSRG6VE56ZZVd6k7Y104FhkoZI6gqcDUwqKvN74BhJdZJ2Ao4CXpLUXVLPJBR1B04EXmzrgG7Smlk2ZXoeXkQ0SroImALUAhMiYqakcen68RHxkqT7gOeBDcC1EfGipH2AO5Uk1Trgpoi4r61jOuGZWUblu7UsIiYDk4uWjS+avxy4vGjZPNKmbRZOeGaWnZ+HZ2a5UaW3ljnhmVk28tNSzCxPXMMzs7yQE56Z5UHyhHcnPDPLAwnVOOGZWU64hmdmueGEZ2a54YRnZvkgmn+wUxVwwjOzTIRcwzOz/Kip8Z0WZpYTruGZWT74HJ6Z5YlreGaWC+60MLNc8a1lZpYPcpPWzHLECc/McqNaE151Xj1oZp2mqdOilFeb+5LGSpojaa6kS1soM0bSs5JmSvpTlm2LuYZnZtmVoYInqRa4CjgBqAemS5oUEbMKyvQGfgmMjYiFkvqVum1zXMMzs2yU3FpWyqsNI4G5ETEvItYBE4FTi8p8CrgjIhYCRMTSDNtuwQnPzDLL0KTtK+mpgtcFBbvZC1hUMF+fLiu0L9BH0lRJMySdm2HbLbhJa2bZld6kbYiIERn2EkXzdcARwPFAN+AJSX8pcdstOOFtg+NH7c8Pv/Zxamtq+O3vH+eK3zywRZnRhw/jh187k7q6Wt5YuZpTLvxvAL5wznF85rSjIYJZcxfzpctu4N11jR39FnLnj4/P4ls/u431GzbwmVOP5l/OO3Gz9ZP/9DzfH38PNRJ1dTX84KsfZ9RhQ6l/bQVf+O71LF3+JjUSnz19NOPOOa6T3kXnK1MvbT0woGB+b2BxM2UaImINsEbSI8ChJW67hXZNeJLGAv8N1ALXRsSP2vN4HammRlz+zU9y+kVXsvj1lTz0m2/wh0deYM4rr20ss3OPbvz0kk/yiYt/Sf3rK+jbpwcAe+zWiwvP+hAfOOv7vPPue0z4wfmcceIR3HzPk531dnJh/foNfOMnt3LnlRex5+69+fBnL+ekYw9mv3322Fjm2COHc9KxByOJF19+lfO/NYFpt32buroa/usrZ3DofgN4a807HHfujxlz1H6bbZsXpfbAlmA6MEzSEOBV4GySc3aFfg9cKakO6AocBfwcmF3Ctltot3N4Bb0oJwEHAOdIOqC9jtfRjjhwMPMWNbDg1eW817ieOx54mpM/dMhmZT4xdgT3PPwc9a+vAKBhxeqN6+rqatlxhy7U1taw045deW3Zqg6NP49mzJzPPgP6MnjvvnTtUscZJxzO5D89v1mZHjvtsPGX+e21724cb7p/314cul9SoejZfUf2HdyfJctWdmT4FaUcl6VERCNwETAFeAm4NSJmShonaVxa5iXgPuB5YBpJxenFlrZtK+72rOFt7EUBkNTUi9Jqt3G12GO3XryaJjKAxa+v4IiDBm9WZujAfnSpq+Xu8f9Mj512YPzEqdwyeRpLlq3iFzc8yAt3f4933l3Hw0/O5uEnZ3fwO8ifJctWsdfufTbO77l7H2a8OH+Lcvc8/ByXXTWJZSve4pafj9ti/cLFy3l+Tj1HHDi4HaOtbOW6lzYiJgOTi5aNL5q/HLi8lG3b0p69tCX1oki6oKkHJxrXtmM45dXcX68oOmVaV1vDofsN4Kyv/C9nfvkqvvG5sQwd2I9ePbtx8rEHc9ip32H/k/6NnXbsyidPOrKDIs+vKP6CgOYqIaccdyjTbvs2N1x+AT8Yf+9m61a//S7nXnItP/zqmezco1t7hVrxynXhcUdrz4RXUi9KRFwdESMiYoTqqucHaPHSlVvUFl5rWLVFmQefeIm331nHG6vW8Pgzczlo2F6MGbkfCxYvZ/nK1TSu38DdDz/HyEOGdPRbyJ09+/Xeolbev2+vFsuPPvx9vPJqA8tXJqci3mtcz2cvuYZPjB3BRz98WHuHW7nkhNecrepFqRZPz1rA0IG7MXDPXelSV8sZJxzOHx7Z/HzQ5D89z6j3D6W2toZuO3RhxEGD+ev816h/7Q1GHDyEbjt0AeBDRw5nziuvd8bbyJXDDxjE3xYuY8GrDax7r5E7Hniak47d/LzrvEXLNtYEn5u9iPfea2SXXt2JCL78vRvZd3B/vvTp4zsj/IohkppxKa9K057n8Erpgala69dv4Js/uZXb/+dL1NaKGyf9hdnzXuMfz/ggAP93x2P8df7rPPj4LB676VtEBNf//nFe+tsSACY9+AxTb7iE9es38Pycen5z55878+3kQl1dLT/55ic58+KrWL8++PTHPsD+Q/dgwu2PAnD+mccw6aFnueXeJ6mrq6Xbjl349Q/ORxJPPPs3bpk8jQPetyfHfOqHAHz7Sx/jxNEHduZb6iSVWXsrhZo7r1G2nUsnA1eQXJYyISK+31r5mp36xQ7DP9lu8Vj5rZh+ZWeHYBmMPmoEM2Y8tU3Zasf++8agz/6ipLJ//cnYGa1ceNzh2vU6vK3pRTGzClehzdVS+E4LM8tEJBfeVyMnPDPLzDU8M8uNau20cMIzs2x8Ds/M8kKolId7ViQnPDPLzDU8M8sNn8Mzs3zwOTwzy4vkXtrqzHhOeGaWWZXmOyc8M8vOd1qYWT7ITVozy4mm5+FVIyc8M8uoep+H54RnZplVab5r10e8m9n2SEmnRSmvNncljZU0R9JcSZc2s36MpFWSnk1f/1Gwbr6kF9LlT5USumt4ZpZJua7DKxi7+gSSMXCmS5oUEcVDuT4aEae0sJvjIqKh1GO6hmdmmZVp1LKNY1dHxDqgaezqduOEZ2aZZRi1rG/TuNPp64KC3ZQ0djUwStJzkv4gqXDUpADulzSjaL8tcpPWzDLL0KRtaGUQn1LGrn4aGBQRq9NBwe4ChqXrRkfEYkn9gAckzY6IR1oLxjU8M8umxNpdCTmxzbGrI+LNiFidTk8Gukjqm84vTv9fCtxJ0kRulROemWWSPAC0LL20G8eultSVZOzqSZsdS+qvtDopaSRJzlouqbuknuny7sCJwIttHdBNWjPLrKYMvbQR0SjpImAKm8aunilpXLp+PPBx4AuSGoG1wNkREZJ2B+5Mc2EdcFNE3NfWMZ3wzCyzcl143NzY1Wmia5q+EthitPeImAccmvV4Tnhmlon88AAzy5MqfTpUywlP0i/Ysot4o4i4uF0iMrOKtz0+D6+ke9PMLF9E0lNbjVpMeBHxm8J5Sd0jYk37h2Rmla5KK3htX4cnaZSkWcBL6fyhkn7Z7pGZWWUq8T7aSuzYKOXC4yuAvwOWA0TEc8Cx7RiTmVW4Mt1p0eFK6qWNiEVF2Xp9+4RjZpVOlOfC485QSsJbJOloINLbPy4mbd6aWT5Vay9tKU3accCXSB7b8ipwWDpvZjlUanO2EiuBbdbw0qeJfroDYjGzKlGtTdpSemn3kXS3pGWSlkr6vaR9OiI4M6tMKvFVaUpp0t4E3ArsAewJ/A64uT2DMrPKtj1flqKI+G1ENKavG2jlljMz274lvbSlvSpNa/fS7pJOPpwOnzaRJNGdBdzbAbGZWSVSaUMwVqLWOi1mkCS4pnd2YcG6AL7XXkGZWWWrxOZqKVq7l3ZIRwZiZtWhqUlbjUq600LSQcABwI5NyyLi+vYKyswq23ZXw2si6TvAGJKENxk4CXgMcMIzy6nqTHel9dJ+HDgeeC0i/pHkOfI7tGtUZlaxJKitUUmvSlNKk3ZtRGyQ1ChpZ2Ap4AuPzXKsWpu0pdTwnpLUG7iGpOf2aWBaewZlZpWtXPfSShoraY6kuenlb8Xrx0haJenZ9PUfpW7bnFLupf1iOjle0n3AzhHxfCk7N7Ptj1BZ7qWVVAtcBZwA1APTJU2KiFlFRR+NiFO2ctvNtHbh8eGtrYuIp1t9N2a2fSrfk1BGAnPTMWaRNBE4FWg1aW3Ltq3V8H7WyroAPlxCUJkMGNifS676erl3a+3o9Gue7OwQLIO5DeUZlibDOby+kgoHBLs6Iq5Op/cCFhWsqweOamYfoyQ9BywGvh4RMzNsu5nWLjw+rq2NzSx/BNSWnvAaImJEK7sqVnyf/tPAoIhYLelk4C5gWInbbqGUTgszs82U6eEB9cCAgvm9SWpxG0XEmxGxOp2eDHSR1LeUbZuNu82QzMyKlCnhTQeGSRqSDh9xNjCpsICk/krbz5JGkuSs5aVs25ySbi0zM2uSXHKy7b0WEdEo6SJgClALTIiImZLGpevHk9z48AVJjcBa4OyICKDZbds6Zim3lonkEe/7RMRlkgYC/SPC1+KZ5VS5bqJIm6mTi5aNL5i+Eriy1G3bUkqT9pfAKOCcdP4tkutfzCyntttBfICjIuJwSc8ARMSKtM1sZjkkoK4Ss1kJSkl476VXNQeApN2ADe0alZlVtCrNdyUlvP8B7gT6Sfo+yUnEf2/XqMysYknlubWsM5RyL+2NkmaQPCJKwGkR8VK7R2ZmFatK811JvbQDgbeBuwuXRcTC9gzMzCpXBT7qriSlNGnvZdNgPjsCQ4A5wIHtGJeZVShBRT7csxSlNGkPLpxPn6JyYQvFzWx7V6FjzpYi850WEfG0pCPbIxgzqw6q0lEtSjmH99WC2RrgcGBZu0VkZhVtex+msWfBdCPJOb3b2yccM6sG22XCSy847hER3+igeMysClTrID6tPeK9Ln2aQYuPejez/EmGaezsKLZOazW8aSTn656VNAn4HbDx+dARcUc7x2ZmFWq7vdMC2IXkgXsfZtP1eAE44Znl0PbaadEv7aF9kU2Jrkmbz443s+1XlVbwWk14tUAPtnKwDDPbXoma7fA6vCURcVmHRWJmVUFsnzW8Kn1LZtauBHVVehKvtYR3fIdFYWZVY7us4UXEGx0ZiJlVj2q9LKVKLx80s85UrkF8JI2VNEfSXEmXtlLuSEnrJX28YNl8SS9IelbSU6XE7XFpzSwTUZ6aUnrr6lXACUA9MF3SpIiY1Uy5H5OMQVvsuIhoKPWYruGZWTZKmrSlvNowEpgbEfMiYh0wETi1mXJfJnlgydJtDd0Jz8wySe60KDnh9ZX0VMHrgoJd7QUsKpivT5dtOpa0F3A6MJ4tBXC/pBlF+22Rm7RmllmGLouGiBiRYTfFNzVcAVwSEeubeULL6IhYLKkf8ICk2RHxSGvBOOGZWWZl6qStBwYUzO8NLC4qMwKYmCa7vsDJkhoj4q6IWAwQEUsl3UnSRG414blJa2YZCam0VxumA8MkDZHUFTgbmFRYICKGRMTgiBgM3AZ8MSLuktRdUk8ASd2BE0nu+2+Va3hmlkm5emnT521eRNL7WgtMiIiZksal65s7b9dkd+DONKnWATdFxH1tHdMJz8wyK9eFxxExGZhctKzZRBcR5xVMzwMOzXo8Jzwzy0bb4SPezcyaU64mbWdwwjOzzFzDM7PcqM5054RnZhkJqHUNz8zyokrznROemWUlVKWNWic8M8vMNTwzy4XkspTqzHhOeGaWTYlPM65ETnhmllm1jmnhhGdmmSQPAO3sKLaOE56ZZeZeWjPLjSpt0TrhbYvZM1/hrtseYsOG4KjRB3P8iUc1W27hgiX8z+U38ZnzT+HQw4cDsPbtd7j1xiksWbIcAWf9w1gG77NnB0afT+/fuxefGzWIGok/zlnKHc8t2Wz9gXv05Fsn7svSt94F4C+vrODWZ17duL5GcPlpB/HG2+v4/pS/dmjslcQ1vCKSJgCnAEsj4qD2Ok5n2bBhA3fc+kcu/PIn6NW7J1f85AYOPHgo/ffou0W5e+96hOH7D95s+V23PcTwA4bw2X86lcbG9by37r0OjD6fagQXjB7MdyfPZvmadfzktAOZtmAl9SvXblbupdfeajGZnXJQf+pXrmWnrrUdEXJFquZzeO35lJfrgLHtuP9OtXD+a+y6Wx927duburpa3n/Efsx8/m9blHts6jMcfNi+9Oi508Zl76x9l3lz6znq6IMBqKurpdtOO3ZY7Hk1bLceLHnzHV5/610aNwSP/e0NRg7qU/L2u3bvyhEDevPHOcvaMcoqUOKIZZXYk9tuCS8dPeiN9tp/Z1u18i169+m5cb5X7x6sWvnWFmVeeO5ljj5m8wezLm9YRfceOzHxt/fxsx9ezy03TuHdd9d1SNx5tkv3rjSs3vQ5L1+zjl27d9mi3PB+Pfh/ZxzEt8cOZ0CfbhuXn/+BQfxm2kI2RPHAWvmjEl+VptOf4yfpgqYxK1evXN7Z4Wybor9od932MKecdiw1NZt/zBs2bODVRa9z9DGH8bVvncsOXbvw0P3TOjLSXCplTMB5DW9zwc3P8tU7XuTema9x6Qn7AjBiYG9WvfMe8xrebvc4K13GcWkrSqd3WkTE1cDVAIP2O6Rq/nT26t2TlSs21ehWrVxNr149NitTv/A1fjvhHgDWrF7L7JnzqKmtYdDgPenVuyeDhuwBwCHv35eH7n+y44LPqeVr1tG3R9eN87t278obazY/d7r2vfUbp59etIoLR4ueO9Sx3+49OXJgH44Y0JsutWKnrrV8ZcxQrpi65WmMPKi8VFaaTk941WrAoP40LF3B8oaV9Ordk2dmzOYfzvv7zcr822WbBkO/+fo/cMBB+3DwocMA6N2nJ0tff4N+u+/Cy3MWsHv/XTs0/jx6edlq9th5R/r13IE31qzjg0N34ecPb56wenfrwsq1SRIctlt3JHjr3UZumL6IG6YvApKe3NMO2SO3yQ6o2oznhLeVamtrOOOTx3P1VbcTGzYwctTB9N+zL48/+iwARx9zWKvbn/6J47nxuntZ37ieXfr25uzPbLf9OxVjQ8A1j8/nOycNp0biwTnLWLRiLX+3fz8Apry0lFFDdmHsAf1YvyFY1xj87MG5nRx1ZSpXc1XSWOC/SYZpvDYiftRCuSOBvwBnRcRtWbbdbD/RTidgJd0MjCEZLfx14DsR8evWthm03yFxyYRJrRWxCnP3c693dgiWweM/Oo9VC17apmy1/8Hvj+t/P7WksiOH9p4RESOaWyepFvgrcAJQTzIw9zkRMauZcg8A75CMXXtbqdsWa7caXkSc0177NrNOVp4K3khgbjrGLJImAqcCxUnry8DtwJFbse1mOr2X1syqS3LJSWn/gL5NV2GkrwsKdrUXsKhgvj5dtulY0l7A6UDx4Nxtbtscn8Mzs2yyPQ+voaUmLaVdKXQFcElErC8aGrKUbbfghGdmmZWpk7YeGFAwvzewuKjMCGBimuz6AidLaixx2y044ZlZRirXQNzTgWGShgCvAmcDnyosEBFDNh5Vug64JyLuklTX1rbNccIzs8zKke8iolHSRcAUkktLJkTETEnj0vXF5+3a3LatYzrhmVkm5bxPNiImA5OLljWb6CLivLa2bYsTnpll5zstzCwv/ABQM8uNCnwQSkmc8MwsG49La2Z54iatmeWCcA3PzHKkSvOdE56ZbYUqzXhOeGaWWSWOV1EKJzwzy6w6050TnpltjSrNeE54ZpZJ0wNAq5ETnpll4wuPzSxPqjTfOeGZWVZlewBoh3PCM7PMqjTfOeGZWTblfABoR3PCM7PsqjTjOeGZWWa+LMXMcsPn8MwsHwQ1VZrwajo7ADOrRirx1cZepLGS5kiaK+nSZtafKul5Sc9KekrSBwvWzZf0QtO6UqJ2Dc/MMinXA0Al1QJXAScA9cB0SZMiYlZBsQeBSRERkg4BbgX2K1h/XEQ0lHpM1/DMLLPy1O8YCcyNiHkRsQ6YCJxaWCAiVkdEpLPdgWAbOOGZWWZSaS+gb9oUbXpdULCbvYBFBfP16bKiY+l0SbOBe4HzC1YFcL+kGUX7bZGbtGaWWYZbyxoiYkRLu2lm2RY1uIi4E7hT0rHA94CPpKtGR8RiSf2AByTNjohHWgvGNTwzy6xMTdp6YEDB/N7A4pYKp8lsqKS+6fzi9P+lwJ0kTeRWOeGZWSalNmdLqAROB4ZJGiKpK3A2MGnzY+l9SquTkg4HugLLJXWX1DNd3h04EXixrQO6SWtmmZXjTouIaJR0ETAFqAUmRMRMSePS9eOBM4FzJb0HrAXOSntsdydp5kKSx26KiPvaOqYTnpllV6YLjyNiMjC5aNn4gukfAz9uZrt5wKFZj+eEZ2aZVemNFk54ZpaVPEyjmeVDue606AzupTWz3HANz8wyq9YanhOemWXmB4CaWT54XFozy4tq7rRwwjOzzNykNbPccA3PzHKjSvOdE56ZbYUqzXhOeGaWiaBqby3TpsfFdz5Jy4AFnR1HO+gLlDzQiFWE7fU7GxQRu23LDiTdR/L5lKIhIsZuy/HKqaIS3vZK0lOtPObaKpC/s+2T76U1s9xwwjOz3HDC6xhXd3YAlpm/s+2Qz+GZWW64hmdmueGEZ2a54YTXjiSNlTRH0lxJl3Z2PNY2SRMkLZXU5hinVn2c8NqJpFrgKuAk4ADgHEkHdG5UVoLrgIq5UNbKywmv/YwE5kbEvIhYB0wETu3kmKwNEfEI8EZnx2Htwwmv/ewFLCqYr0+XmVknccJrP83dXe1rgMw6kRNe+6kHBhTM7w0s7qRYzAwnvPY0HRgmaYikrsDZwKROjsks15zw2klENAIXAVOAl4BbI2Jm50ZlbZF0M/AEMFxSvaTPdXZMVj6+tczMcsM1PDPLDSc8M8sNJzwzyw0nPDPLDSc8M8sNJ7wqImm9pGclvSjpd5J22oZ9XSfp4+n0ta092EDSGElHb8Ux5kvaYnSrlpYXlVmd8VjflfT1rDFavjjhVZe1EXFYRBwErAPGFa5Mn9CSWUR8PiJmtVJkDJA54ZlVGie86vUo8L609vWwpJuAFyTVSrpc0nRJz0u6EECJKyXNknQv0K9pR5KmShqRTo+V9LSk5yQ9KGkwSWL9l7R2eYyk3STdnh5juqTR6ba7Srpf0jOSfkUJ49NLukvSDEkzJV1QtO5naSwPStotXTZU0n3pNo9K2q8sn6blQl1nB2DZSaojec7efemikcBBEfFKmjRWRcSRknYA/izpfuD9wHDgYGB3YBYwoWi/uwHXAMem+9olIt6QNB5YHRE/TcvdBPw8Ih6TNJDkbpL9ge8Aj0XEZZL+HtgsgbXg/PQY3YDpkm6PiOVAd+DpiPiapP9I930RyeA64yLiZUlHAb8EPrwVH6PlkBNedekm6dl0+lHg1yRNzWkR8Uq6/ETgkKbzc0AvYBhwLHBzRKwHFkt6qJn9fwB4pGlfEdHSc+E+AhwgbazA7SypZ3qMM9Jt75W0ooT3dLGk09PpAWmsy4ENwC3p8huAOyT1SN/v7wqOvUMJxzADnPCqzdqIOKxwQfqLv6ZwEfDliJhSVO5k2n48lUooA8mpkFERsbaZWEq+V1HSGJLkOSoi3pY0FdixheKRHndl8WdgViqfw9v+TAG+IKkLgKR9JXUHHgHOTs/x7QEc18y2TwAfkjQk3XaXdPlbQM+CcveTNC9Jyx2WTj4CfDpddhLQp41YewEr0mS3H0kNs0kN0FRL/RRJU/lN4BVJn0iPIUmHtnEMs42c8LY/15Kcn3s6HYjmVyQ1+TuBl4EXgP8F/lS8YUQsIznvdoek59jUpLwbOL2p0wK4GBiRdorMYlNv8X8Cx0p6mqRpvbCNWO8D6iQ9D3wP+EvBujXAgZJmkJyjuyxd/mngc2l8M/Fj8y0DPy3FzHLDNTwzyw0nPDPLDSc8M8sNJzwzyw0nPDPLDSc8M8sNJzwzy43/DyzneLbiIOeCAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "class_names = [0,1]\n",
    "\n",
    "classifier = grid_search\n",
    "\n",
    "np.set_printoptions(precision=2)\n",
    "\n",
    "# Plot non-normalized confusion matrix\n",
    "titles_options = [(\"Confusion matrix, without normalization\", None),\n",
    "                  (\"Normalized confusion matrix\", 'true')]\n",
    "for title, normalize in titles_options:\n",
    "    disp = plot_confusion_matrix(classifier, X_test, y_test,\n",
    "                                 display_labels=class_names,\n",
    "                                 cmap=plt.cm.Blues,\n",
    "                                 normalize=normalize)\n",
    "    disp.ax_.set_title(title)\n",
    "\n",
    "    print(title)\n",
    "    print(disp.confusion_matrix)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      0.68      0.78     17606\n",
      "           1       0.18      0.54      0.27      2263\n",
      "\n",
      "    accuracy                           0.67     19869\n",
      "   macro avg       0.55      0.61      0.53     19869\n",
      "weighted avg       0.84      0.67      0.72     19869\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.36      0.52     17606\n",
      "           1       0.14      0.81      0.24      2263\n",
      "\n",
      "    accuracy                           0.41     19869\n",
      "   macro avg       0.54      0.58      0.38     19869\n",
      "weighted avg       0.84      0.41      0.49     19869\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "\n",
    "target_names = y_test.unique().astype(str)\n",
    "probs_y = grid_search.predict_proba(X_test)\n",
    "y_pred = grid_search.best_estimator_.predict(X_test)\n",
    "print(classification_report(y_test, y_pred, target_names=target_names))\n",
    "\n",
    "vals = []\n",
    "\n",
    "for val in probs_y[:,1]:\n",
    "    if val > 0.4:\n",
    "        vals.append(1)\n",
    "    else:\n",
    "        vals.append(0)\n",
    "\n",
    "print(classification_report(y_test, vals, target_names=target_names))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot ROC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.roc_curve.RocCurveDisplay at 0x7ff31ff02510>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "mt.plot_roc_curve(grid_search.best_estimator_, X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.65 0.35]\n",
      " [0.28 0.72]\n",
      " [0.68 0.32]\n",
      " ...\n",
      " [0.74 0.26]\n",
      " [0.63 0.37]\n",
      " [0.62 0.38]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([0.35, 0.72, 0.32, ..., 0.26, 0.37, 0.38])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "probs_y = grid_search.predict_proba(X_test)\n",
    "print(probs_y)\n",
    "\n",
    "probs_y[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.0, 1.0)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, probs_y[:,1]) \n",
    "\n",
    "plt.title(\"Precision-Recall vs Threshold Chart\")\n",
    "plt.plot(thresholds, precision[: -1], \"b--\", label=\"Precision\")\n",
    "plt.plot(thresholds, recall[: -1], \"r--\", label=\"Recall\")\n",
    "plt.ylabel(\"Precision, Recall\")\n",
    "plt.xlabel(\"Threshold\")\n",
    "plt.legend(loc=\"lower left\")\n",
    "plt.ylim([0,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Threshold: 0.3155047469923995\n",
      "Precision: 0.12107901109421637\n",
      "Recall: 0.9500662836942112\n"
     ]
    }
   ],
   "source": [
    "readmit95_ind = np.argmin(recall >= 0.95)-1\n",
    "readmit95_thresh = thresholds[readmit95_ind]\n",
    "readmit95_precision = precision[readmit95_ind]\n",
    "readmit95_recall = recall[readmit95_ind]\n",
    "\n",
    "print(\"Threshold:\", readmit95_thresh)\n",
    "print(\"Precision:\", readmit95_precision)\n",
    "print(\"Recall:\", readmit95_recall)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(19852,)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Proportion of samples below threshold: 0.648873096242312\n"
     ]
    }
   ],
   "source": [
    "print(\"Proportion of samples below threshold:\", np.mean(grid_search.best_estimator_.predict(X) < readmit95_thresh))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/fabiosavorgnan/opt/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# classifier\n",
    "classifier = grid_search.fit(X_train, y_train)\n",
    "y_pred_proba = classifier.predict_proba(X_test)[::,1]\n",
    "fpr, tpr, _ = roc_curve(y_test,  y_pred_proba)\n",
    "auc = roc_auc_score(y_test, y_pred_proba)\n",
    "plt.plot(fpr,tpr,label=\"AUC=\"+str(auc))\n",
    "plt.legend(loc=4)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tunning the threshold for classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "******** For i = 0.01 ******\n",
      "0.11389601892395189\n",
      "Precision= 0.11389601892395189\n",
      "Recal= 0.11389601892395189\n",
      "fscore= 0.023291766747235054\n",
      "Confusion Matrix = [[    0 17606]\n",
      " [    0  2263]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "******** For i = 0.1 ******\n",
      "0.11389601892395189\n",
      "Precision= 0.11389601892395189\n",
      "Recal= 0.11389601892395189\n",
      "fscore= 0.023291766747235054\n",
      "Confusion Matrix = [[    0 17606]\n",
      " [    0  2263]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/d3fzzAAAACXBIWXMAAAsTAAALEwEAmpwYAAAlDklEQVR4nO3deViVdf7/8edHBHFBXHAFEdx3TVGyTTPLpdJMpxpLyyxrZpxmmt+kVDpZOmU1822aaTGzmrbRJsA1lzbbs9RSNkURN8QVFBBkPZ/fHzhdZChHPXA457we18Ul97k/5z7vD8vLm/vc9/s21lpERMTz1XF3ASIi4hoKdBERL6FAFxHxEgp0EREvoUAXEfESdd31wiEhITYiIsJdLy8i4pE2b958zFrborJ1bgv0iIgINm3a5K6XFxHxSMaYvWdbp0MuIiJeQoEuIuIlFOgiIl5CgS4i4iUU6CIiXqLKQDfGvG6MOWKMSTrLemOM+acxJs0Yk2CM6e/6MkVEpCrO7KH/Gxh5jvWjgM6nP6YBL198WSIicr6qDHRr7RdA9jmGjAXesuU2AE2MMW1cVaCIiLcoKXPw0mdpbN1/olq274pj6KHA/grLGacf+wVjzDRjzCZjzKajR4+64KVFRDxD0oEcbnrxa55Zm8qapEPV8hquuFLUVPJYpXfNsNYuBBYCREVF6c4aIuL1CkvK+NenO1nweTpNGwTw8u39GdW7eg5iuCLQM4B2FZbDgEwXbFdExKNt2pPNjLgE0o/m86sBYcy6vgfBDfyr7fVcEegrgOnGmCVANJBjrT3ogu2KiHikk0WlPLt2O29t2Evb4Pq8dfcgrupSaT8tl6oy0I0xi4GhQIgxJgN4DPAHsNYuAFYDo4E0oACYUl3FiojUdp/vOMoj8Ylk5pzizsERPDSiKw3r1UwfxCpfxVr76yrWW+B3LqtIRMQDnSgoZu6qbcT9kEHHFg15/77BREU0q9Ea3NY+V0TEW6xJPMjs5ckcLyhm+tWdmD6sE4H+fjVehwJdROQCHckt5C/Lk1mbfIheoY158+6B9Gwb7LZ6FOgiIufJWsv7mzOYtyqFwlIHM0d2494rI6nr5972WAp0EZHzsD+7gEeWJvLlzmMMimjG/PG96dCikbvLAhToIiJOKXNY3vp2D8+uS8UAc8f25Pbo9tSpU9m1le6hQBcRqULakTxmxiWyee9xhnRpwZM39ya0SX13l/ULCnQRkbMoKXPwyue7+OcnaTSo58dzt/blpn6hGFN79sorUqCLiFQiMSOHGXEJbDuYy/V92vD4mJ6ENKrn7rLOSYEuIlJBYUkZ//h4J69+mU7zhgG8MmkAI3q2dndZTlGgi4ic9l16FjHxiew+ls+tUe145PruBNevvmZarqZAFxGfl1dYwjNrU3l7w17aNavPu/dEc3mnEHeXdd4U6CLi09anHuHR+EQO5hZy9+WR/HlEFxoEeGY0embVIiIX6Xh+MXNXpRD/4wE6t2xE3G8uo394U3eXdVEU6CLiU6y1fJB4kMeWJ5NzqoQHrunM767uSL26Nd9My9UU6CLiMw7nFjJrWRIfpRymT1gw79wTTfc2jd1dlsso0EXE61lr+e+m/cz7YBvFpQ4eGd2Nuy93fzMtV1Ogi4hX25dVQEx8At/syiI6shlPj+9DREhDd5dVLRToIuKVyhyWf3+zh7+tS8WvjuGv43rx64HhtaqZlqsp0EXE6+w4nMeM2AS27D/BsG4t+eu4XrQJrn3NtFxNgS4iXqO41MHLn+3ihfU7CQr05/nb+jGmb9ta20zL1RToIuIVtu4/wcy4BLYfymNM37Y8dmMPmtfyZlqupkAXEY92qriM5z7ewaIv02kZFMiiyVEM79HK3WW5hQJdRDzWt7uyeDg+gT1ZBfx6UDgPj+5G40DPaablagp0EfE4uYUlzF+znf98t4/2zRvwn3ujuayj5zXTcjUFuoh4lE+2HebRpUkcyStk2lUdeHB4F+oHeP5l+66gQBcRj5B1sojHV6awYmsmXVsFsWDSAPq1a+LusmoVBbqI1GrWWlZszeTxlSnkFZbw4PAu/GZoRwLqetdl+66gQBeRWutgzilmLU3ik+1H6NuuCc+M70PX1kHuLqvWUqCLSK3jcFiWbNzPU6u3UeJwMOv67ky5PBI/L75s3xUU6CJSq+w5lk9MfAIb0rO5rGNznrq5N+2be2czLVdToItIrVBa5uD1r3fz9w93EOBXh/k39+bWge185rJ9V3Aq0I0xI4HnAT9gkbV2/hnrg4F3gPDT2/ybtfYNF9cqIl5q+6FcZsYmsDUjh+HdWzHvpl60Dg50d1kep8pAN8b4AS8C1wIZwEZjzAprbUqFYb8DUqy1NxpjWgCpxph3rbXF1VK1iHiFotIyXly/i5fWpxFc359//foSbujTRnvlF8iZPfRBQJq1Nh3AGLMEGAtUDHQLBJny70IjIBsodXGtIuJFftx3nJlxCew4fJJxl4Qy+4YeNGsY4O6yPJozgR4K7K+wnAFEnzHmBWAFkAkEAbdaax1nbsgYMw2YBhAeHn4h9YqIhysoLuXvH+7g9a9307pxIK/fFcWwbr7ZTMvVnAn0yv72sWcsjwC2AMOAjsBHxpgvrbW5P3uStQuBhQBRUVFnbkNEvNw3aceIiU9kX3YBd1wazsyR3Qjy4WZaruZMoGcA7Sosh1G+J17RFGC+tdYCacaY3UA34HuXVCkiHi3nVAlPrd7Gko37iQxpyHvTLiW6Q3N3l+V1nAn0jUBnY0wkcAC4DZh4xph9wDXAl8aYVkBXIN2VhYqIZ/ow+RCzliVx7GQR9w0pb6YV6K9mWtWhykC31pYaY6YD6yg/bfF1a22yMeb+0+sXAHOBfxtjEik/RDPTWnusGusWkVru2Mki5qxIZlXCQbq1DmLRnVH0CWvi7rK8mlPnoVtrVwOrz3hsQYXPM4HrXFuaiHgiay3Lthzg8ZUpFBSV8f+u7cL9Qzvi76dmWtVNV4qKiMtknjjFo0sTWZ96lEvCy5tpdW6lZlo1RYEuIhfN4bC8+/0+5q/ehsPCYzf2YPLgCDXTqmEKdBG5KOlHTxITl8j3e7K5olMIT93cm3bNGri7LJ+kQBeRC1Ja5mDRV7t57qMd1Ktbh2cm9OFXA8J02b4bKdBF5LylZOYyI24rSQdyGdGzFXPH9qJlYzXTcjcFuog4rai0jBc+TePlz3bRpIE/L93en1G9WmuvvJZQoIuIUzbvzWZmXCJpR04yvn8Ys67vTlM106pVFOgick75RaU8uy6VN7/dQ9vg+rx59yCGdGnh7rKkEgp0ETmrL3ce5eH4RDKOn+LOwe15aGQ3GtVTbNRW+s6IyC/kFJQw74MU3t+cQYcWDXn//sEMjGjm7rKkCgp0EfmZtUmHmL08iez8Yn47tCMPXNNZzbQ8hAJdRAA4klfInBXJrE48RI82jXnjroH0Cg12d1lyHhToIj7OWkvcDweYuyqFUyVlPDSiK9Ou6qBmWh5IgS7iwzKOF/DI0iS+2HGUqPZNmT++D51aNnJ3WXKBFOgiPsjhsLy9YS9Pr90OwONjejLp0vbUUTMtj6ZAF/Exu46eZGZsApv2HueqLi14clwvwpqqmZY3UKCL+IiSMgcLv0jn+U92Ut/fj7/9qi/j+4fqsn0vokAX8QFJB3KYEZtAysFcRvduzZwxPWkZpGZa3kaBLuLFCkvKeP6TnSz8Ip1mDQNYcEd/RvZq4+6ypJoo0EW81MY92cyMTSD9WD6/GhDGrOt7ENzA391lSTVSoIt4mZNFpTyzdjtvfbuXsKb1eXvqIK7srGZavkCBLuJFPt9xlEfiE8nMOcVdl0Xw0IiuNFQzLZ+h77SIFzhRUMwTq1KI/+EAHVs0JPb+wQxor2ZavkaBLuLBrLWsSTrEX5YncaKghOlXd2L6sE5qpuWjFOgiHupIbiGzlyexLvkwvUIb8+bdg+jZVs20fJkCXcTDWGt5f3MG81alUFTqIGZUN+65IpK6aqbl8xToIh5kf3YBD8cn8lXaMQZFNGP++N50aKFmWlJOgS7iAcoclre+3cMza1OpY2DuTb24fVC4mmnJzyjQRWq5nYfzmBmXwA/7TjC0awv+Oq43oU3qu7ssqYUU6CK1VEmZgwWf7eJfn6bRsJ4fz93al5v6qZmWnJ1TgW6MGQk8D/gBi6y18ysZMxT4B+APHLPWDnFZlSI+JjEjh4dit7L9UB439GnDnDE9CWlUz91lSS1XZaAbY/yAF4FrgQxgozFmhbU2pcKYJsBLwEhr7T5jTMtqqlfEqxWWlPHcxzt49Yt0QhrVY+GkAVzXs7W7yxIP4cwe+iAgzVqbDmCMWQKMBVIqjJkIxFtr9wFYa4+4ulARb/ddehYx8YnsPpbPbQPb8fDo7gTXVzMtcZ4zgR4K7K+wnAFEnzGmC+BvjPkMCAKet9a+deaGjDHTgGkA4eHhF1KviNfJKyzh6bXbeWfDPto1q8+790RzeacQd5clHsiZQK/sHRhbyXYGANcA9YFvjTEbrLU7fvYkaxcCCwGioqLO3IaIz1m//QiPLE3kUG4hU6+I5P9d14UGATpXQS6MMz85GUC7CsthQGYlY45Za/OBfGPMF0BfYAci8gvZ+cU8sTKZZVsy6dyyEXG/uYz+4U3dXZZ4OGcCfSPQ2RgTCRwAbqP8mHlFy4EXjDF1gQDKD8k858pCRbyBtZZVCQeZsyKZnFMl/OGazvz26o7Uq6tmWnLxqgx0a22pMWY6sI7y0xZft9YmG2PuP71+gbV2mzFmLZAAOCg/tTGpOgsX8TSHcwt5dGkSH287TJ+wYN69N5purRu7uyzxIsZa9xzKjoqKsps2bXLLa4vUJGst723cz19Xb6O41MGfr+vKlMsj1ExLLogxZrO1NqqydXr3RaQa7c3K5+H4RL7ZlUV0ZDOeHt+HiJCG7i5LvJQCXaQalDksb3y9m799mErdOnV4clxvbhvYTs20pFop0EVcLPVQHjPiEti6/wTXdGvJvHG9aBOsZlpS/RToIi5SXOrgpc/SeHF9GkGB/jx/Wz/G9G2rZlpSYxToIi6wdf8JZsQmkHo4j7H92vKXG3rQXM20pIYp0EUuwqniMv7vo1Re+2o3LYMCWTQ5iuE9Wrm7LPFRCnSRC/TNrmM8HJ/I3qwCJkaHEzOqG40D1UxL3EeBLnKecgtLeGr1dhZ/v4/2zRvwn3ujuayjmmmJ+ynQRc7DxymHeXRZIkfziph2VQceHN6F+gG6bF9qBwW6iBOyThbx+MoUVmzNpFvrIBZOiqJvuybuLkvkZxToIudgrWXF1kzmrEjmZFEpDw7vwm+GdiSgri7bl9pHgS5yFgdzTjFraRKfbD9Cv3ZNeGZCH7q0CnJ3WSJnpUAXOYPDYVm8cR9Prd5OqcPBrOu7M+XySPx02b7Ucgp0kQp2H8snJi6B73Znc1nH5sy/uQ/hzRu4uywRpyjQRYDSMgevf72bv3+4g4C6dXh6fG9uiWqny/bFoyjQxedtO5jLzLgEEjJyuLZHK+bd1ItWjQPdXZbIeVOgi88qKi3jxfW7eGl9GsH1/Xlh4iVc37uN9srFYynQxSf9sO84M2MT2HnkJOMuCeUvN/SgacMAd5clclEU6OJTCopL+du6HbzxzW5aNw7kjbsGcnW3lu4uS8QlFOjiM75OO0ZMfAL7s08x6dL2zBjZlSA10xIvokAXr5dzqoQnP9jGe5v2ExnSkPemXUp0h+buLkvE5RTo4tU+TD7ErGVJZOUXc/+QjvxxeGcC/dVMS7yTAl280tG8IuasTOaDhIN0b9OY1+4cSO+wYHeXJVKtFOjiVay1LP3xAE+sSqGgqIw/X9eF+4Z0xN9PzbTE+ynQxWscOHGKR5cm8lnqUfqHlzfT6tRSzbTEdyjQxeM5HJZ3v9vL/DXbcVh47MYeTB4coWZa4nMU6OLR0o+eJCYuke/3ZHNl5xCeHNebds3UTEt8kwJdPFJpmYNXv9zNcx/vILBuHZ6d0IcJA8J02b74NAW6eJzkzBxmxiWQdCCXET1bMXdsL1qqmZaIAl08R2FJGf/6dCcLPk+naYMAXr69P6N6t3F3WSK1hgJdPMLmvdnMiE1g19F8xvcPY/YN3WnSQM20RCpy6uRcY8xIY0yqMSbNGBNzjnEDjTFlxpgJritRfFl+USlzViQzYcG3FJY4ePPuQfz9lr4Kc5FKVLmHbozxA14ErgUygI3GmBXW2pRKxj0NrKuOQsX3fLHjKA/HJ5KZc4rJl7bnoZHdaFRPf1SKnI0zvx2DgDRrbTqAMWYJMBZIOWPc74E4YKBLKxSfk1NQwtwPUojdnEGHFg35732DGRjRzN1lidR6zgR6KLC/wnIGEF1xgDEmFBgHDOMcgW6MmQZMAwgPDz/fWsUHrE06yOzlyWTnF/PboR154Bo10xJxljOBXtmJvfaM5X8AM621Zec6D9hauxBYCBAVFXXmNsSHHckr5LHlyaxJOkSPNo15466B9ApVMy2R8+FMoGcA7SoshwGZZ4yJApacDvMQYLQxptRau8wVRYr3stYSuzmDeR9s41RJGTNGduXeKzuomZbIBXAm0DcCnY0xkcAB4DZgYsUB1trI/31ujPk3sEphLlXZn13AI0sT+XLnMQZGNGX++D50bNHI3WWJeKwqA91aW2qMmU752St+wOvW2mRjzP2n1y+o5hrFyzgclre+3cMz61IxwBNje3JHdHvqqJmWyEVx6hwwa+1qYPUZj1Ua5Nbauy6+LPFWaUdOEhOXwKa9x7mqSwueHNeLsKZqpiXiCjqpV2pESZmDhV+k8/zHO6kf4Mfff9WXm/uHqpmWiAsp0KXaJR3IYUZsAikHcxnduzWPj+lFi6B67i5LxOso0KXaFJaU8fwnO1n4RTrNGgaw4I4BjOzV2t1liXgtBbpUi417spkZm0D6sXxuiQrj0dE9CG7g7+6yRLyaAl1c6mRRKc+s3c5b3+4lrGl93pkazRWdQ9xdlohPUKCLy6xPPcKj8YkczC1kyuUR/Pm6rjRUMy2RGqPfNrlox/OLmbsqhfgfD9CpZSNi77+MAe2burssEZ+jQJcLZq1ldeIhHluRxImCEn4/rBPTh3WiXl010xJxBwW6XJAjuYXMWpbEhymH6R0azFt3R9OjbWN3lyXi0xTocl6stby/KYO5H6RQXOrg4VHdmHpFJHXVTEvE7RTo4rT92QU8HJ/IV2nHGBTZjPk396aDmmmJ1BoKdKlSmcPy5jd7eHZdKn51DPNu6sXEQeFqpiVSyyjQ5Zx2Hs5jRlwCP+47wdCuLXhyXG/aNqnv7rJEpBIKdKlUcamDBZ/v4oVP02hYz49/3NqPsf3aqpmWSC2mQJdfSMg4wYzYBLYfyuPGvm157MYehDRSMy2R2k6BLj8pLCnjuY928OqX6bQIqserk6O4tkcrd5clIk5SoAsAG9KziIlLYE9WAb8e1I6YUd0Jrq9mWiKeRIHu4/IKS5i/ZjvvfreP8GYN+M890VzWSc20RDyRAt2Hfbr9MI8uTeJwbiH3XBHJn67rQoMA/UiIeCr99vqg7PxinliZzLItmXRp1YiXbr+MS8LVTEvE0ynQfYi1lpUJB5mzIpm8whL+cE1nfnd1JwLq6rJ9EW+gQPcRh3LKm2l9vO0wfcOCeXpCNN1aq5mWiDdRoHs5ay1LNu7nyQ+2UeJw8Ojo7tx9RSR+umxfxOso0L3Y3qx8YuIS+TY9i0s7NGP+zX2ICGno7rJEpJoo0L1QmcPyxte7+duHqfjXqcOT43pz28B2aqYl4uUU6F4m9VB5M62t+09wTbeWzBvXizbBaqYl4gsU6F6iuNTBS5+l8eL6NIIC/fnnry/hxj5t1ExLxIco0L3Alv0nmBmbQOrhPMb2a8tjN/akWcMAd5clIjVMge7BThWX8fcPU3n96920DArktTujuKa7mmmJ+CoFuof6ZtcxYuIS2ZddwMTocGJGdaNxoJppifgypy4RNMaMNMakGmPSjDExlay/3RiTcPrjG2NMX9eXKgC5hSU8HJ/AxFe/wxhYfO+lPDmut8JcRKreQzfG+AEvAtcCGcBGY8wKa21KhWG7gSHW2uPGmFHAQiC6Ogr2ZR+nHObRZYkczSvivqs68MfhXagf4OfuskSklnDmkMsgIM1amw5gjFkCjAV+CnRr7TcVxm8AwlxZpK/LOlnEnJUprNyaSbfWQbw6OYo+YU3cXZaI1DLOBHoosL/Ccgbn3vueCqypbIUxZhowDSA8PNzJEn2XtZblWzJ5fGUyJ4tK+dO1Xbh/SEc10xKRSjkT6JWdyGwrHWjM1ZQH+hWVrbfWLqT8cAxRUVGVbkPKZZ44xaxlSXy6/Qj92jXhmQl96NIqyN1liUgt5kygZwDtKiyHAZlnDjLG9AEWAaOstVmuKc/3OByW/3y/j/lrtlPmsMy+oQd3XRahZloiUiVnAn0j0NkYEwkcAG4DJlYcYIwJB+KBSdbaHS6v0kfsPpZPTFwC3+3O5vJOzXlqXB/Cmzdwd1ki4iGqDHRrbakxZjqwDvADXrfWJhtj7j+9fgHwF6A58NLpS81LrbVR1Ve2dyktc/DaV7v5v492EFC3Ds+M78OvosJ02b6InBdjrXsOZUdFRdlNmza55bVrk5TMXGbGJZB4IIdre7Ri3k29aNU40N1liUgtZYzZfLYdZl0p6iZFpWW88GkaL3+2iyYN/HlxYn9G926tvXIRuWAKdDfYvPc4M+MSSDtykpsvCWX2DT1oqmZaInKRFOg1qKC4lGfXpfLvb/bQpnEgb0wZyNVdW7q7LBHxEgr0GvLVzmPExCeQcfwUky5tz4yRXQlS/xURcSEFejXLOVXCXz9I4b+bMogMach/7xvMoMhm7i5LRLyQAr0arUs+xOxlSWTlF/OboR35wzWdCfRXMy0RqR4K9GpwNK+IOSuS+SDxIN3bNOa1OwfSOyzY3WWJiJdToLuQtZb4Hw7wxKoUThWX8dCIrky7qgP+fmqmJSLVT4HuIgdOnOKR+EQ+33GU/uHlzbQ6tVQzLRGpOQr0i+RwWN75bi9Pr9mOBebc2INJg9VMS0RqngL9Iuw6epKYuAQ27jnOlZ1DeHJcb9o1UzMtEXEPBfoFKClz8OqX6fzj450E1q3DsxP6MGGAmmmJiHsp0M9T0oEcZsYlkJyZy8ierXnipp60DFIzLRFxPwW6kwpLyvjXpztZ8Hk6TRsE8PLt/RnVu427yxIR+YkC3Qmb9mQzIy6B9KP5jO8fxuwbutOkgZppiUjtokA/h/yi8mZab367h7bB9Xnz7kEM6dLC3WWJiFRKgX4Wn+84yiPxiWTmnOLOwRE8NKIrDevpyyUitZcS6gwnCoqZu2obcT9k0KFFQ96/bzBREWqmJSK1nwK9gjWJB5m9PJnjBcX87uqO/H6YmmmJiOdQoANHcgv5y/Jk1iYfomfbxrx590B6tlUzLRHxLD4d6NZaYjdnMHdVCoWlDmaO7MY9V0aqmZaIeCSfDfT92QU8sjSRL3ceY2BEU+aP70PHFo3cXZaI1ygpKSEjI4PCwkJ3l+KRAgMDCQsLw9/f+Tub+Vyglzksb3+7h2fWpWKAuWN7cnt0e+qomZaIS2VkZBAUFERERITaYpwnay1ZWVlkZGQQGRnp9PN8KtDTjuQxMy6RzXuPM6RLC/46rhdhTdVMS6Q6FBYWKswvkDGG5s2bc/To0fN6nk8EekmZg1c+38U/P0mjQT0//u+Wvoy7JFQ/aCLVTL9jF+5CvnZeH+hJB3J4KDaBbQdzub53G+aM6UmLoHruLktExOW89nSOwpIy5q/ZztgXv+bYySIW3DGAF2/vrzAX8TFLly7FGMP27dsB+Oyzz7jhhht+Nuauu+4iNjYWKH8zNyYmhs6dO9OrVy8GDRrEmjVrnHqtoqIibr31Vjp16kR0dDR79uypdNzQoUPp2rUr/fr1o1+/fhw5cuTCJ1iBVwb697uzGf38lyz4fBcT+ofx8YNDGNmrtbvLEhE3WLx4MVdccQVLlixxavzs2bM5ePAgSUlJJCUlsXLlSvLy8px67muvvUbTpk1JS0vjwQcfZObMmWcd++6777Jlyxa2bNlCy5Ytndp+VbzqkEteYQnPrE3l7Q17CWtan3emRnNF5xB3lyXi8x5fmUxKZq5Lt9mjbWMeu7HnOcecPHmSr7/+mvXr1zNmzBjmzJlzzvEFBQW8+uqr7N69m3r1yv+ab9WqFbfccotTNS1fvvyn15gwYQLTp0/HWltj7yV4TaCvTz3Co/GJHMwt5O7LI/nziC40CPCa6YnIBVi2bBkjR46kS5cuNGvWjB9++OGc49PS0ggPD6dx48aVrr/11ltJTU39xeN/+tOfmDx5MgcOHKBdu3YA1K1bl+DgYLKysggJ+eWO5ZQpU/Dz82P8+PHMmjXLJaHv8Yl3PL+YuatSiP/xAJ1aNiL2/ssY0L6pu8sSkQqq2pOuLosXL+aPf/wjALfddhuLFy/+xfHz/3EmUN97771zrrfWOrXdd999l9DQUPLy8hg/fjxvv/02kydPrvL1q+JUoBtjRgLPA37AImvt/DPWm9PrRwMFwF3W2nP/V3iRrLV8kHiQx5Ynk3OqhAeGdeJ3wzpRr66aaYkIZGVl8emnn5KUlIQxhrKyMowxTJ48mePHj/9sbHZ2NiEhIXTq1Il9+/aRl5dHUFDQL7ZZ1R56WFgY+/fvJywsjNLSUnJycmjW7JfdWkNDQwEICgpi4sSJfP/99zUT6MYYP+BF4FogA9hojFlhrU2pMGwU0Pn0RzTw8ul/q8Xh3EJmL0viw5TD9A4N5p17ounepvI/kUTEN8XGxjJ58mReeeWVnx4bMmQI2dnZZGZmsm3bNrp3787evXvZunUr/fr1o0GDBkydOpUHHniAV155hYCAAA4ePMgnn3zCHXfcUeUe+pgxY3jzzTcZPHgwsbGxDBs27Bd76KWlpZw4cYKQkBBKSkpYtWoVw4cPd8mcndlDHwSkWWvTAYwxS4CxQMVAHwu8Zcv/3thgjGlijGljrT3okiorWL/9CA8s+ZHiUgcPj+rG1CsiqatmWiJyhsWLFxMTE/Ozx8aPH8+SJUt45513mDJlCoWFhfj7+7No0SKCg8s7rM6bN49Zs2bRo0cPAgMDadiwIU888YRTrzl16lQmTZpEp06daNas2c/OrOnXrx9btmyhqKiIESNGUFJSQllZGcOHD+fee+91yZxNZcd8fjbAmAnASGvtPaeXJwHR1trpFcasAuZba786vfwJMNNau+mMbU0DpgGEh4cP2Lt373kXvOdYPo+tSGbOmJ5EhjQ87+eLSM343x6wXLjKvobGmM3W2qjKxjuza1vZOwVn/i/gzBistQuttVHW2qgWLS7s3pwRIQ158+5BCnMRkTM4E+gZQLsKy2FA5gWMERGRauRMoG8EOhtjIo0xAcBtwIozxqwAJptylwI51XH8XEQ8S1WHdOXsLuRrV+WbotbaUmPMdGAd5actvm6tTTbG3H96/QJgNeWnLKZRftrilPOuRES8SmBgIFlZWTRv3lxdF8/T//qhBwYGntfzqnxTtLpERUXZTZs2VT1QRDyS7lh0cc52x6JzvSnq8VeKikjt5O/vf15325GLpxO4RUS8hAJdRMRLKNBFRLyE294UNcYcBc7/UtFyIcAxF5bjCTRn36A5+4aLmXN7a22lV2a6LdAvhjFm09ne5fVWmrNv0Jx9Q3XNWYdcRES8hAJdRMRLeGqgL3R3AW6gOfsGzdk3VMucPfIYuoiI/JKn7qGLiMgZFOgiIl6iVge6MWakMSbVGJNmjImpZL0xxvzz9PoEY0x/d9TpSk7M+fbTc00wxnxjjOnrjjpdqao5Vxg30BhTdvouWh7NmTkbY4YaY7YYY5KNMZ/XdI2u5sTPdrAxZqUxZuvpOXt011ZjzOvGmCPGmKSzrHd9fllra+UH5a16dwEdgABgK9DjjDGjgTWU3zHpUuA7d9ddA3O+DGh6+vNRvjDnCuM+pbxV8wR3110D3+cmlN+3N/z0ckt3110Dc34EePr05y2AbCDA3bVfxJyvAvoDSWdZ7/L8qs176D/dnNpaWwz87+bUFf10c2pr7QagiTGmTU0X6kJVztla+4219vjpxQ2U3x3KkznzfQb4PRAHHKnJ4qqJM3OeCMRba/cBWGs9fd7OzNkCQaa8eXojygO9tGbLdB1r7ReUz+FsXJ5ftTnQQ4H9FZYzTj92vmM8yfnOZyrl/8N7sirnbIwJBcYBC2qwrurkzPe5C9DUGPOZMWazMWZyjVVXPZyZ8wtAd8pvX5kI/MFa66iZ8tzC5flVm/uhu+zm1B7E6fkYY66mPNCvqNaKqp8zc/4HMNNaW+Yld75xZs51gQHANUB94FtjzAZr7Y7qLq6aODPnEcAWYBjQEfjIGPOltTa3mmtzF5fnV20OdF+8ObVT8zHG9AEWAaOstVk1VFt1cWbOUcCS02EeAow2xpRaa5fVSIWu5+zP9jFrbT6Qb4z5AugLeGqgOzPnKcB8W36AOc0YsxvoBnxfMyXWOJfnV20+5OKLN6eucs7GmHAgHpjkwXtrFVU5Z2ttpLU2wlobAcQCv/XgMAfnfraXA1caY+oaYxoA0cC2Gq7TlZyZ8z7K/yLBGNMK6Aqk12iVNcvl+VVr99CtD96c2sk5/wVoDrx0eo+11Hpwpzon5+xVnJmztXabMWYtkAA4gEXW2kpPf/METn6f5wL/NsYkUn44Yqa11mPb6hpjFgNDgRBjTAbwGOAP1ZdfuvRfRMRL1OZDLiIich4U6CIiXkKBLiLiJRToIiJeQoEuIuIlFOgiIl5CgS4i4iX+P5SSVjr9bzU5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "******** For i = 0.2 ******\n",
      "0.11746942473199457\n",
      "Precision= 0.11746942473199457\n",
      "Recal= 0.11746942473199457\n",
      "fscore= 0.030849612188469076\n",
      "Confusion Matrix = [[   75 17531]\n",
      " [    4  2259]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "******** For i = 0.25 ******\n",
      "0.13473249786098948\n",
      "Precision= 0.13473249786098948\n",
      "Recal= 0.13473249786098948\n",
      "fscore= 0.06615734732176949\n",
      "Confusion Matrix = [[  434 17172]\n",
      " [   20  2243]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "******** For i = 0.3 ******\n",
      "0.18390457496602747\n",
      "Precision= 0.18390457496602747\n",
      "Recal= 0.18390457496602747\n",
      "fscore= 0.16017375568506537\n",
      "Confusion Matrix = [[ 1470 16136]\n",
      " [   79  2184]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred_proba_df = pd.DataFrame(grid_search.predict_proba(X_test))\n",
    "threshold_list = [0.01, 0.1,0.2, 0.25, 0.3]\n",
    "for i in threshold_list:\n",
    "    print ('\\n******** For i = {} ******'.format(i))\n",
    "    Y_test_pred = pred_proba_df.applymap(lambda x: 1 if x>i else 0)\n",
    "    accuracy = mt.accuracy_score(y_test, Y_test_pred.iloc[:,1])\n",
    "    print(accuracy)\n",
    "    precision = mt.precision_score(y_test, Y_test_pred.iloc[:,1], average='micro')\n",
    "    print(\"Precision=\", precision)\n",
    "    recal = mt.recall_score(y_test, Y_test_pred.iloc[:,1], average='micro')\n",
    "    print(\"Recal=\", recal)\n",
    "    fscore = mt.f1_score(y_test, Y_test_pred.iloc[:,1], average='weighted')\n",
    "    print(\"fscore=\",fscore )\n",
    "    # Confusion matrix\n",
    "    cm= confusion_matrix(y_test, Y_test_pred.iloc[:,1])\n",
    "    print(\"Confusion Matrix =\", cm)\n",
    "    fpr, tpr, _ = roc_curve(y_test, Y_test_pred.iloc[:,1])\n",
    "    auc = roc_auc_score(y_test, Y_test_pred.iloc[:,1])\n",
    "    plt.plot(fpr,tpr,label=\"AUC=\"+str(auc))\n",
    "    plt.legend(loc=4)\n",
    "    plt.show()\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "937ee2fabb34e01a2bfda78bec8fb2c2f970bd2ee382315ee43d2303ee1eab74"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}