Feature_Engineering_V1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Engineering Version 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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",
      "encounter_id                101766 non-null int64\n",
      "patient_nbr                 101766 non-null int64\n",
      "race                        101766 non-null object\n",
      "gender                      101766 non-null object\n",
      "age                         101766 non-null object\n",
      "weight                      101766 non-null object\n",
      "admission_type_id           101766 non-null int64\n",
      "discharge_disposition_id    101766 non-null int64\n",
      "admission_source_id         101766 non-null int64\n",
      "time_in_hospital            101766 non-null int64\n",
      "payer_code                  101766 non-null object\n",
      "medical_specialty           101766 non-null object\n",
      "num_lab_procedures          101766 non-null int64\n",
      "num_procedures              101766 non-null int64\n",
      "num_medications             101766 non-null int64\n",
      "number_outpatient           101766 non-null int64\n",
      "number_emergency            101766 non-null int64\n",
      "number_inpatient            101766 non-null int64\n",
      "diag_1                      101766 non-null object\n",
      "diag_2                      101766 non-null object\n",
      "diag_3                      101766 non-null object\n",
      "number_diagnoses            101766 non-null int64\n",
      "max_glu_serum               101766 non-null object\n",
      "A1Cresult                   101766 non-null object\n",
      "metformin                   101766 non-null object\n",
      "repaglinide                 101766 non-null object\n",
      "nateglinide                 101766 non-null object\n",
      "chlorpropamide              101766 non-null object\n",
      "glimepiride                 101766 non-null object\n",
      "acetohexamide               101766 non-null object\n",
      "glipizide                   101766 non-null object\n",
      "glyburide                   101766 non-null object\n",
      "tolbutamide                 101766 non-null object\n",
      "pioglitazone                101766 non-null object\n",
      "rosiglitazone               101766 non-null object\n",
      "acarbose                    101766 non-null object\n",
      "miglitol                    101766 non-null object\n",
      "troglitazone                101766 non-null object\n",
      "tolazamide                  101766 non-null object\n",
      "examide                     101766 non-null object\n",
      "citoglipton                 101766 non-null object\n",
      "insulin                     101766 non-null object\n",
      "glyburide-metformin         101766 non-null object\n",
      "glipizide-metformin         101766 non-null object\n",
      "glimepiride-pioglitazone    101766 non-null object\n",
      "metformin-rosiglitazone     101766 non-null object\n",
      "metformin-pioglitazone      101766 non-null object\n",
      "change                      101766 non-null object\n",
      "diabetesMed                 101766 non-null object\n",
      "readmitted                  101766 non-null object\n",
      "dtypes: int64(13), object(37)\n",
      "memory usage: 38.8+ MB\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "df = pd.read_csv(\"diabetic_data.csv\")\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Caucasian          76099\n",
       "AfricanAmerican    19210\n",
       "?                   2273\n",
       "Hispanic            2037\n",
       "Other               1506\n",
       "Asian                641\n",
       "Name: race, dtype: int64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['race'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "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",
      "encounter_id                101766 non-null int64\n",
      "patient_nbr                 101766 non-null int64\n",
      "race                        99493 non-null object\n",
      "gender                      101766 non-null object\n",
      "age                         101766 non-null object\n",
      "weight                      3197 non-null object\n",
      "admission_type_id           101766 non-null int64\n",
      "discharge_disposition_id    101766 non-null int64\n",
      "admission_source_id         101766 non-null int64\n",
      "time_in_hospital            101766 non-null int64\n",
      "payer_code                  61510 non-null object\n",
      "medical_specialty           51817 non-null object\n",
      "num_lab_procedures          101766 non-null int64\n",
      "num_procedures              101766 non-null int64\n",
      "num_medications             101766 non-null int64\n",
      "number_outpatient           101766 non-null int64\n",
      "number_emergency            101766 non-null int64\n",
      "number_inpatient            101766 non-null int64\n",
      "diag_1                      101745 non-null object\n",
      "diag_2                      101408 non-null object\n",
      "diag_3                      100343 non-null object\n",
      "number_diagnoses            101766 non-null int64\n",
      "max_glu_serum               101766 non-null object\n",
      "A1Cresult                   101766 non-null object\n",
      "metformin                   101766 non-null object\n",
      "repaglinide                 101766 non-null object\n",
      "nateglinide                 101766 non-null object\n",
      "chlorpropamide              101766 non-null object\n",
      "glimepiride                 101766 non-null object\n",
      "acetohexamide               101766 non-null object\n",
      "glipizide                   101766 non-null object\n",
      "glyburide                   101766 non-null object\n",
      "tolbutamide                 101766 non-null object\n",
      "pioglitazone                101766 non-null object\n",
      "rosiglitazone               101766 non-null object\n",
      "acarbose                    101766 non-null object\n",
      "miglitol                    101766 non-null object\n",
      "troglitazone                101766 non-null object\n",
      "tolazamide                  101766 non-null object\n",
      "examide                     101766 non-null object\n",
      "citoglipton                 101766 non-null object\n",
      "insulin                     101766 non-null object\n",
      "glyburide-metformin         101766 non-null object\n",
      "glipizide-metformin         101766 non-null object\n",
      "glimepiride-pioglitazone    101766 non-null object\n",
      "metformin-rosiglitazone     101766 non-null object\n",
      "metformin-pioglitazone      101766 non-null object\n",
      "change                      101766 non-null object\n",
      "diabetesMed                 101766 non-null object\n",
      "readmitted                  101766 non-null object\n",
      "dtypes: int64(13), object(37)\n",
      "memory usage: 38.8+ MB\n"
     ]
    }
   ],
   "source": [
    "# Replace ? with NaN, as pandas has built-in functions to handle NaN\n",
    "df.replace('?',np.nan,inplace=True)\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(10,8))\n",
    "missing = pd.DataFrame({'column':df.columns ,'na_percent':df.isnull().sum()/len(df)*100})\n",
    "missing.sort_values('na_percent',inplace=True)\n",
    "plt.barh(missing['column'],width=missing['na_percent'])\n",
    "plt.title(\"Percentage Missing Values in columns\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df[df['gender']=='Unknown/Invalid'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "encounter_id                0\n",
       "patient_nbr                 0\n",
       "race                        0\n",
       "gender                      0\n",
       "age                         0\n",
       "admission_type_id           0\n",
       "discharge_disposition_id    0\n",
       "admission_source_id         0\n",
       "time_in_hospital            0\n",
       "num_lab_procedures          0\n",
       "num_procedures              0\n",
       "num_medications             0\n",
       "number_outpatient           0\n",
       "number_emergency            0\n",
       "number_inpatient            0\n",
       "diag_1                      0\n",
       "diag_2                      0\n",
       "diag_3                      0\n",
       "number_diagnoses            0\n",
       "max_glu_serum               0\n",
       "A1Cresult                   0\n",
       "metformin                   0\n",
       "repaglinide                 0\n",
       "nateglinide                 0\n",
       "chlorpropamide              0\n",
       "glimepiride                 0\n",
       "acetohexamide               0\n",
       "glipizide                   0\n",
       "glyburide                   0\n",
       "tolbutamide                 0\n",
       "pioglitazone                0\n",
       "rosiglitazone               0\n",
       "acarbose                    0\n",
       "miglitol                    0\n",
       "troglitazone                0\n",
       "tolazamide                  0\n",
       "examide                     0\n",
       "citoglipton                 0\n",
       "insulin                     0\n",
       "glyburide-metformin         0\n",
       "glipizide-metformin         0\n",
       "glimepiride-pioglitazone    0\n",
       "metformin-rosiglitazone     0\n",
       "metformin-pioglitazone      0\n",
       "change                      0\n",
       "diabetesMed                 0\n",
       "readmitted                  0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#dropping columns with high percentage of missing values\n",
    "df.drop(['weight','medical_specialty','payer_code'],axis=1,inplace=True)\n",
    "# Replace uninformative gender with missing value\n",
    "df.loc[df['gender'] == 'Unknown/Invalid', 'gender'] = np.nan\n",
    "# dropping all records which contain NaN\n",
    "df.dropna(inplace=True)\n",
    "df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(98052, 47)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "68629"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['patient_nbr'].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "88785891     39\n",
       "1660293      23\n",
       "88227540     23\n",
       "23199021     23\n",
       "84428613     22\n",
       "             ..\n",
       "36674964      1\n",
       "106407315     1\n",
       "78154461      1\n",
       "5461389       1\n",
       "93251151      1\n",
       "Name: patient_nbr, Length: 68629, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['patient_nbr'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the paper: \"The preliminary dataset contained multiple inpatient\n",
    "visits for some patients and the observations could not\n",
    "be considered as statistically independent, an assumption\n",
    "of the logistic regression model. We thus used only one\n",
    "encounter per patient; in particular, we considered only the\n",
    "first encounter for each patient as the primary admission and\n",
    "determined whether or not they were readmitted within 30\n",
    "days.\"\n",
    "\n",
    "This would be an issue for logistic our regression. Hence, we drop all the repeated occurrences of patients to make the data i.i.d.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Keep only the first encounter of each patient\n",
    "df['total_admissions'] = df.groupby('patient_nbr').cumcount()\n",
    "df = df.drop_duplicates(subset= ['patient_nbr'], keep = 'last')\n",
    "# Drop encounter_id and patient_nbr\n",
    "df.drop(['encounter_id','patient_nbr'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At this point of time, the dataset seemed to be ready for one-hot encoding. But applying it produced ~2350 columns. We need a further look into each of the columns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "race ['Caucasian' 'AfricanAmerican' 'Other' 'Asian' 'Hispanic']\n",
      "gender ['Female' 'Male']\n",
      "age ['[10-20)' '[20-30)' '[30-40)' '[40-50)' '[50-60)' '[60-70)' '[70-80)'\n",
      " '[80-90)' '[90-100)' '[0-10)']\n",
      "admission_type_id [1 2 3 6 4 5 8 7]\n",
      "discharge_disposition_id [ 1  3  6  2  5 11 25 10  7 14  4 18  8 13 12 22 17 23  9 16 20 15 28 24\n",
      " 19 27]\n",
      "admission_source_id [ 7  2  4  1  6 20  5  3 17  8  9 14 10 22 11 25 13]\n",
      "time_in_hospital [ 3  2  1  4  5 13 12  9  7 10 11  6  8 14]\n",
      "num_lab_procedures [ 59  11  44  51  31  70  73  68  33  47  62  55  49  75  45  29  35  42\n",
      "  19  64  52  87  53  46  28  36  48  72  10   2  65  67  58  37  57  25\n",
      "  34  39  69  38  27  56  22  96  78  61  60  66  43   1  18  41  82  54\n",
      "   9  24  71  63  77  50  81  93  76 103  80  90  16  32  12  30  17  21\n",
      "  79  26  95  97  14  74  23  13  40 105  86  83  84  98  20   5  85   6\n",
      "  94   8 102 100  15  88  91  92   4  89 101   3  99   7 114 111 129 107\n",
      " 108 106 109 120 113 104 121 126 118]\n",
      "num_procedures [0 5 1 6 2 3 4]\n",
      "num_medications [18 13 16  8 21 12 28 17 11 31  2 23 19  7 14 10 20 22 15  9 27 25  6 30\n",
      " 26  5  4 24  3 61  1 46 33 41 36 32 29 50 43 42 37 34 51 35 38 45 40 54\n",
      " 52 49 39 47 44 48 57 53 59 62 56 63 55 58 64 69 65 66 60 67 68 81 79 75\n",
      " 74]\n",
      "number_outpatient [ 0  2  1  5  3  9  8  4 12 11  6 20 15  7 10 14 35 13 16 17 29 36 18 19\n",
      " 22 21 26 25 24 27 37 40 33]\n",
      "number_emergency [ 0  1  2  4  9  6  3  8  7 25 10  5 42 16 11 28 15 14 12 37 19 13 20 18\n",
      " 54 29]\n",
      "number_inpatient [ 0  1  2  3  6  5  4  7  8  9 15 10 11 13 12 16]\n",
      "diag_1 ['276' '648' '8' '197' '414' '428' '398' '434' '250.7' '157' '518' '999'\n",
      " '410' '682' '402' 'V57' '189' '786' '427' '584' '462' '174' '998' '996'\n",
      " '511' '432' '626' '295' '250.6' '182' '845' '423' '808' '722' '403' '784'\n",
      " '707' '440' '151' '715' '997' '198' '564' '590' '486' '411' '578'\n",
      " '250.32' '433' '569' '185' '536' '250.4' '255' '599' '558' '574' '491'\n",
      " '560' '244' '730' '188' '824' '250.8' '332' '562' '296' '510' '70' '642'\n",
      " '250.03' '571' '593' '618' '456' '446' '250.02' '820' '515' '780'\n",
      " '250.22' '38' '577' '995' '235' '196' '250.82' '721' '556' '162' '724'\n",
      " '401' '250.83' 'V55' '281' '466' '530' '789' '566' '822' '250.13' '277'\n",
      " '711' '575' '202' '493' '280' '553' '441' '349' '557' '962' '507' '386'\n",
      " '156' '200' '728' '435' '426' '337' '82' '250.12' '225' '596' '288' '656'\n",
      " '787' '492' '220' '516' '210' '922' '572' '286' '885' '958' '661' '969'\n",
      " '227' '404' '823' '191' '416' '535' '250.41' '250' '211' '218' '457'\n",
      " '303' '431' '550' '351' '601' '723' '204' '424' '153' '241' '358'\n",
      " '250.81' '345' '681' '447' '290' '158' '436' '335' '453' '812' '309'\n",
      " '654' '805' '799' '292' '455' '250.42' '482' '733' 'V58' '78' '851' '532'\n",
      " '531' '738' '586' '311' '54' '892' '415' '591' '794' '803' '250.11' '79'\n",
      " '293' '540' '429' '278' '658' '592' '443' '729' '444' '727' '154' '214'\n",
      " '458' '291' '284' '585' '644' '481' '821' '413' '595' '437' '285' '968'\n",
      " '756' '340' '275' '512' '781' '420' '737' '825' '304' '320' '250.31'\n",
      " '669' '868' '3' '680' '459' '53' '438' '495' '49' '346' '294' '852'\n",
      " '250.43' '576' '850' '287' '331' '250.2' '611' '782' '840' '350' '726'\n",
      " '250.33' '620' '180' '783' '598' '751' '250.3' '199' '580' '836' '664'\n",
      " '283' '359' '496' '289' '708' '573' '184' '608' '333' '212' '117' '788'\n",
      " '625' '924' '621' '785' '714' '942' '250.23' '47' '552' '567' '250.1'\n",
      " '844' '7' '233' '42' '397' '421' '807' '253' '600' '604' '39' '659' '312'\n",
      " '473' '614' '652' '274' '861' '425' '527' '201' '451' '348' '485'\n",
      " '250.53' '970' '193' 'V53' '160' '322' '581' '475' '716' '582' '568'\n",
      " '465' '238' '801' '237' '537' '376' '461' '913' '251' '959' '987' '641'\n",
      " '480' '790' '336' '362' '228' '513' '478' '383' '746' '353' '911' '298'\n",
      " '522' '155' '442' '860' '534' '802' '141' 'V45' '310' '341' '242' '305'\n",
      " '736' '533' '616' '203' '519' '301' 'V66' '5' '965' '989' '230' '300'\n",
      " '853' '871' '848' '112' '934' '250.21' '361' '719' '810' '9' '430' '528'\n",
      " '205' '791' '983' '396' '172' '643' '171' '622' '863' '282' '474' '607'\n",
      " '759' '356' '239' '634' '967' '555' '551' '187' '732' '747' '323' '594'\n",
      " '308' '152' '250.92' '183' '846' '164' '250.93' '365' '718' '11' '48'\n",
      " '720' '565' '150' '797' '813' '170' '878' '904' '490' '882' '843' '709'\n",
      " '973' '454' '939' '487' '229' '617' '991' '250.5' '226' '483' '357' '692'\n",
      " '796' '693' '800' '873' '920' '261' '831' '145' '388' '223' 'V71' '839'\n",
      " '685' '710' '395' 'V54' '992' '655' '663' '464' '179' '964' 'V56' '324'\n",
      " '389' '815' '334' '143' '526' '307' '494' '588' '250.01' '977' '192'\n",
      " '135' 'V67' '394' '917' '88' '219' '325' '501' '792' '717' '994' '990'\n",
      " '793' '380' '207' '637' '195' '847' '827' '31' '306' '686' '891' '814'\n",
      " 'V60' '161' '703' '865' '352' '627' '368' '660' '378' '342' '886' '344'\n",
      " '745' '297' '705' '694' '541' '986' '633' '640' '369' '173' '753' '835'\n",
      " '379' '355' '945' '619' '881' '370' '866' '405' '916' '215' '75' '514'\n",
      " '816' '252' '972' '897' '725' '263' '867' '115' '734' '674' '521' '623'\n",
      " '470' '834' '524' '691' '508' '445' '384' '142' '262' '879' '250.51'\n",
      " '246' '695' '208' '893' '646' '955' '653' '149' '245' '146' '966' '570'\n",
      " '854' '933' '35' '136' '194' '163' '579' '216' '374' '94' '477' '318'\n",
      " '880' '921' '377' '935' '683' '448' '175' '602' '250.91' '982' '382'\n",
      " '706' '417' '647' '131' '250.9' '463' '385' '735' '870' '936' '148' 'V26'\n",
      " '862' '272' '61' '817' '610' '914' '360' '684' '314' 'V63' '57' '240'\n",
      " '915' '883' '980' '971' '452' '864' '963' '327' '41' '842' '645' '665'\n",
      " '110' '944' '952' '923' '412' '363' '957' '976' '299' '273' '266' '34'\n",
      " '98' '605' '941' '806' '66' '84' '837' '657' '523' '542' '114' '338'\n",
      " '372' '696' '354' '36' 'E909' 'V07' '826' '422' '731' '279' '919' '875'\n",
      " '236' '890' '147' '381' '632' '804' '23' '58' '795' '529' '649' '615'\n",
      " '832' '583' '928' '271' '133' '975' '833' '391' '690' '250.52' '603' '10'\n",
      " 'V51']\n",
      "diag_2 ['250.01' '250' '250.43' '157' '411' '492' '427' '198' '403' '288' '998'\n",
      " '507' '174' '425' '401' '715' '496' '428' '410' '585' '135' '41' '571'\n",
      " '276' '997' '599' '424' '491' '553' '707' '286' '440' '493' '242' '70'\n",
      " '250.03' '357' 'V45' '511' '196' '396' '197' '414' '250.52' '577' '413'\n",
      " '53' '780' '518' '566' '250.6' '285' '486' '8' '250.02' '788' '340' '228'\n",
      " '250.82' '786' '294' '567' '785' '512' '305' '648' '560' 'V15' '578'\n",
      " '277' '781' '250.42' 'V10' '584' '153' '272' '733' '881' '574' '250.41'\n",
      " '250.13' '293' '999' '250.12' '342' '626' '458' '710' '415' 'V42' '284'\n",
      " '759' '112' '292' '435' '426' '996' '280' '290' '535' '642' '398' '444'\n",
      " '319' '203' '711' '303' '446' '255' 'V44' '250.7' '250.51' '682' '300'\n",
      " '787' '581' '162' '402' '287' '789' '790' '591' '200' '38' '304' '117'\n",
      " '154' '847' '852' '729' '250.11' '816' '575' '416' '412' '515' '482'\n",
      " '456' '250.83' '283' '536' '432' '530' '569' '721' '434' '271' '368'\n",
      " '227' '572' '783' '573' '576' '253' '309' '250.91' '519' '333' '204'\n",
      " '250.92' '250.4' '179' '345' '661' '433' '537' '205' '295' '784' '714'\n",
      " '211' '590' 'E812' '263' '250.23' '244' 'E932' '250.81' '201' '382' '723'\n",
      " '250.5' '722' '34' '861' 'V57' '724' 'E888' '730' '397' '278' '462' '150'\n",
      " '451' '738' 'V43' '155' '202' '218' '220' 'E937' '794' '564' '436' '805'\n",
      " '250.22' '620' '621' '331' '453' '617' '596' '625' '404' '681' '558'\n",
      " '281' '531' '443' '568' '799' '311' 'E878' '719' 'E944' 'V12' '423'\n",
      " 'E870' 'E849' '782' '481' 'V23' '199' '438' '348' '583' '473' '627' '726'\n",
      " '54' '490' '731' '332' '508' '437' 'E939' '79' '208' '922' '562' '431'\n",
      " '296' '258' 'E935' '728' '344' '191' '873' '616' 'V17' '136' '455'\n",
      " '250.8' 'E885' '860' '441' '607' '466' '603' '753' '484' 'V72' '291'\n",
      " '151' 'V58' '510' '185' '592' '791' '598' '362' '217' '250.93' '298'\n",
      " '821' 'E880' '343' '429' '250.53' '716' 'E879' '465' '138' '420' '42'\n",
      " '658' '969' '275' '595' '618' '601' '447' '808' 'E890' '301' '380' '359'\n",
      " '274' '337' '570' 'V14' '324' 'E931' '792' '532' 'V64' '557' '478' '813'\n",
      " '593' '864' 'E934' '718' '459' '386' '807' '604' 'E942' '580' '241' 'V65'\n",
      " '282' '824' '550' '279' '646' '701' '736' '600' '565' '383' '820' '452'\n",
      " '872' '189' '905' '349' '614' '921' '131' '611' '389' '394' '421' '705'\n",
      " '336' '623' '9' '299' '172' '365' '40' '691' '746' '250.31' '745' '250.1'\n",
      " '351' '695' '346' '918' '947' '520' '188' '31' '356' '737' 'V08' '322'\n",
      " '182' '974' 'E929' '378' '252' '94' '680' '702' 'V49' '923' '647' '696'\n",
      " '358' '934' '487' '709' '796' '250.9' '156' '850' '485' '556' '915' '756'\n",
      " '867' '840' '825' '693' '725' '683' '953' '552' '501' '727' 'E900' '812'\n",
      " 'E930' 'V09' '522' '919' '461' '516' '273' '958' '259' '494' '297' '802'\n",
      " 'V54' '320' '906' '239' '608' '454' 'V46' '823' '226' '795' 'E947' '684'\n",
      " '306' '844' '250.33' '308' '615' '457' '712' '663' '706' '833' '713'\n",
      " '533' '372' 'E884' '78' '233' '586' '214' '555' '717' 'E933' '755' 'E928'\n",
      " '869' '962' 'V11' '910' '543' '373' '870' '152' '810' '907' '517' '924'\n",
      " '995' '845' '474' '442' '751' '323' '472' '464' '250.32' '540' '251'\n",
      " '811' '317' '652' '659' '851' '826' '422' '815' '110' '325' '463' '920'\n",
      " '992' '692' 'E950' '521' '579' '524' '741' '193' '173' '238' '335' '137'\n",
      " '405' '514' '528' '225' '312' '837' '355' '594' '622' '289' '369' '475'\n",
      " '965' '500' '754' '261' '891' '868' '968' '381' '250.21' '694' '610'\n",
      " '734' '310' '130' '246' '959' '634' 'E936' 'E927' 'E905' '314' '183'\n",
      " '379' 'E917' '163' 'E868' '495' '747' '989' '430' '240' '832' '644' 'V16'\n",
      " 'V18' '35' '882' 'V70' '801' 'E918' '477' '656' '46' '883' 'V13' '842'\n",
      " 'E850' '800' '269' '75' '664' '171' '952' '164' '836' '250.2' '27' 'V63'\n",
      " '911' '793' '232' '245' '990' '376' '327' '542' '806' '972' '862' 'E829'\n",
      " 'E919' '944' 'E916' '480' '963' '316' 'E858' '513' 'V62' '645' '686'\n",
      " '347' 'V85' '374' '748' '866' '967' '395' '354' '994' 'E826' '641' '685'\n",
      " 'V53' '822' '307' '145' '797' '913' '353' 'E938' 'E816' '948' '892' '99'\n",
      " '192' '250.3' 'E906' '534' '527' '366' '698' '945' 'E814' '266' '758'\n",
      " 'E980' 'E881' 'E941' '360' '338' 'V02' '470' '529' '871' '212' '302'\n",
      " '955' '865' '814' '654' 'E924' 'E819' '750' '88' 'V25' '215' '11' '350'\n",
      " '933' '483' 'V50' 'E968' 'E882' '140' '703' '262' 'E821' '588' 'V69'\n",
      " '831' '670' '195' 'V66' '268' '352' '909' '853' '260' '7' '991' '448'\n",
      " '880' '377' 'V86' '619' '341' '180' '523' '863' '649' 'E853' '270' '665'\n",
      " '460' '942' '364' '66' 'E883' '123' '742' 'V60' 'V55' '843' 'E887' '927']\n",
      "diag_3 ['255' 'V27' '403' '250' 'V45' '38' '486' '996' '197' '427' '627' '414'\n",
      " '416' '582' 'V43' '250.42' '276' '250.41' '250.01' '585' '781' '998'\n",
      " '428' '568' '682' '278' '250.02' '305' '707' '496' '599' '518' '553'\n",
      " '401' '794' '411' 'V42' '511' '490' '562' '250.8' '482' '250.7' '784'\n",
      " '491' '420' '730' '789' '131' '250.82' '999' '424' '41' '493' '250.03'\n",
      " '529' '263' 'E888' '425' '595' '560' '711' '8' '296' '250.4' '654' '244'\n",
      " 'V70' '625' '250.51' 'V10' '280' '440' '581' '569' '272' '997' '250.43'\n",
      " '918' '54' '788' '196' '461' '535' '584' '891' '332' '648' '250.6' '780'\n",
      " '182' '285' '413' '664' '201' 'V15' '292' '473' '714' '564' 'E932' '348'\n",
      " '459' '437' '733' '525' '250.53' '397' '572' '805' '453' '331' '736'\n",
      " '402' '465' '533' '787' '315' '658' '608' '786' '284' '382' '300' '492'\n",
      " '438' '571' '536' '287' '644' '578' 'V11' 'E885' '162' '198' '218' '303'\n",
      " '412' '396' 'V14' '570' '433' 'E934' '882' '458' '288' '357' '785' '577'\n",
      " '443' '729' '836' '295' '281' '153' '410' '616' '799' '304' '250.83'\n",
      " '291' '512' '660' '426' '250.5' '598' '337' '574' '715' '653' '507' '722'\n",
      " '415' '386' '112' '620' '70' '593' '663' '530' '42' '342' '250.91' 'E884'\n",
      " '307' '704' '596' '728' '583' '710' '238' '441' '573' '290' '319'\n",
      " '250.13' '250.12' '346' '380' '642' '250.52' '698' '294' '924' 'E933'\n",
      " '309' 'E879' '286' '753' '752' '731' '681' '444' '311' '344' '455' '35'\n",
      " '558' '790' '813' '394' '301' 'V17' '802' '435' 'V12' '881' '724' '716'\n",
      " 'E935' '250.81' '718' '282' '202' '334' '398' 'V44' 'E849' 'E878' '614'\n",
      " '626' '591' '531' '250.9' '883' '289' '617' '429' '442' 'V25' '866' '610'\n",
      " '557' '959' '94' '920' '362' '345' '79' '293' '516' '365' '586' '821'\n",
      " '600' '242' '373' '592' '250.92' '487' '706' '519' 'E947' '340' '656'\n",
      " 'E949' 'V58' '590' 'V09' '250.22' '694' '138' '250.93' '995' '709' '185'\n",
      " '923' '958' '275' 'E929' '517' 'V18' 'E942' 'V66' '665' '53' '279' '522'\n",
      " '791' '456' 'E938' 'E816' '122' '436' '721' 'V65' '515' '136' '480' '135'\n",
      " '537' '575' '351' '336' '580' '404' '434' '227' '157' '208' 'V57' '466'\n",
      " '274' '734' '150' '199' 'V23' '555' 'E950' '447' 'V64' '228' 'V16' '756'\n",
      " 'E928' '552' '647' '389' '576' '792' '366' '253' '907' '266' '203' '387'\n",
      " '783' '782' '245' '607' 'E930' '705' '372' '611' 'V46' '601' '110' '867'\n",
      " 'E956' '478' '310' '251' '250.2' '712' '695' '343' '746' '567' '723'\n",
      " 'V08' '369' '623' '807' '446' '618' '451' '701' '34' 'V53' '472' '283'\n",
      " '189' '534' '354' '333' 'V54' '659' 'E939' '708' '655' '816' '670' '621'\n",
      " '246' '953' '865' '277' '646' '151' '174' '378' '78' '298' '840' '641'\n",
      " '745' '211' '719' '521' '619' '506' '259' '594' '355' '532' '204' '356'\n",
      " '566' '905' '47' '737' '358' '191' '965' '812' '921' '727' '27' 'E861'\n",
      " '758' '477' '652' '432' '188' '825' '919' '732' '726' '908' '951' '962'\n",
      " '685' '527' '845' '793' '693' '494' '250.1' '273' '860' '155' '692' '579'\n",
      " '969' '205' '462' 'V49' 'E880' '556' '697' '826' '240' '717' '738' '742'\n",
      " '680' '735' '200' '308' '528' '241' '824' '405' '464' '260' '457' '452'\n",
      " '239' '661' '892' '261' '725' 'E883' 'E936' '945' '696' '318' '967'\n",
      " 'E944' '350' '9' '741' 'E941' '170' '317' '759' 'V22' '831' '180' '801'\n",
      " '360' '359' '501' '335' '250.11' '306' '811' '690' 'V02' '271' '214'\n",
      " '873' '847' '543' 'V63' '906' '842' '686' '445' '368' '808' '861' 'E852'\n",
      " '220' 'E887' '423' 'E858' '915' '713' '970' '820' 'E980' '256' '747'\n",
      " '395' '243' '815' '823' '5' '796' '349' '297' '454' '851' '864' '922'\n",
      " '384' 'E931' '270' '225' '158' 'E937' '871' '605' '88' 'E917' 'E812'\n",
      " 'E924' '233' '797' '172' '421' '852' 'E819' '299' '972' 'V72' '916' '956'\n",
      " '703' '173' '193' '154' '347' '862' 'E817' '934' '323' '250.3' '987'\n",
      " '262' '379' '751' 'E855' 'E927' '161' '115' '383' 'V62' '910' '17' '265'\n",
      " 'V60' '870' '944' '844' '417' '152' '183' '252' 'E905' '917' '514' '385'\n",
      " 'E906' '470' '935' '510' '485' 'E870' '912' '312' '850' '179' 'E919'\n",
      " '872' '377' '495' '195' '588' '431' '11' '991' '933' '171' '156' '868'\n",
      " '250.21' 'E815' '542' 'V61' '853' '163' 'E882' '235' 'E822' '810' '754'\n",
      " '523' 'V55' 'V85' '992' '657' 'E881' '604' 'E826' '550' '327' 'E916'\n",
      " '314' 'E943' '460' '376' '313' '755' '361' '720' '250.31' '674' '374'\n",
      " '500' 'E813' '226' 'E818' '955' 'E894' 'E920' '880' '930' 'E915' '381'\n",
      " '132' '353' '795' 'V13' '913' '117' 'E853' '911' '540' '483' 'E828' '430'\n",
      " '800' 'E865' 'E946' 'V86' '338' '388' 'V03' '139' '123' '966' '757'\n",
      " 'E850' '848' '508' 'E892' '649' '702' '948' 'E886' '928' '481' '624'\n",
      " '909' 'E987' '841' 'E955' '217' '877' 'E825' '952' '814' '192' '7' '669'\n",
      " '989' '141' '875' '822' 'E900' 'V06' '215' '980' '834' '448' '175' '49'\n",
      " '876' '230' '57' '216' '258' '268' '943' 'E854' '942' '14' '750' '370'\n",
      " '671' '971' '565']\n",
      "number_diagnoses [ 9  6  7  5  8  3  4 12 13 16 10 11 15 14]\n",
      "max_glu_serum ['None' '>300' 'Norm' '>200']\n",
      "A1Cresult ['None' '>7' '>8' 'Norm']\n",
      "metformin ['No' 'Steady' 'Up' 'Down']\n",
      "repaglinide ['No' 'Steady' 'Up' 'Down']\n",
      "nateglinide ['No' 'Steady' 'Up' 'Down']\n",
      "chlorpropamide ['No' 'Steady' 'Down' 'Up']\n",
      "glimepiride ['No' 'Steady' 'Down' 'Up']\n",
      "acetohexamide ['No' 'Steady']\n",
      "glipizide ['No' 'Steady' 'Up' 'Down']\n",
      "glyburide ['No' 'Steady' 'Up' 'Down']\n",
      "tolbutamide ['No' 'Steady']\n",
      "pioglitazone ['No' 'Steady' 'Up' 'Down']\n",
      "rosiglitazone ['No' 'Steady' 'Up' 'Down']\n",
      "acarbose ['No' 'Steady' 'Up' 'Down']\n",
      "miglitol ['No' 'Steady' 'Up' 'Down']\n",
      "troglitazone ['No' 'Steady']\n",
      "tolazamide ['No' 'Steady']\n",
      "examide ['No']\n",
      "citoglipton ['No']\n",
      "insulin ['Up' 'No' 'Steady' 'Down']\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "glyburide-metformin ['No' 'Steady' 'Down' 'Up']\n",
      "glipizide-metformin ['No' 'Steady']\n",
      "glimepiride-pioglitazone ['No']\n",
      "metformin-rosiglitazone ['No']\n",
      "metformin-pioglitazone ['No' 'Steady']\n",
      "change ['Ch' 'No']\n",
      "diabetesMed ['Yes' 'No']\n",
      "readmitted ['>30' 'NO' '<30']\n",
      "total_admissions [ 0  1  2  4  3  5  6 11  7 14  9 10  8 17 12 15 22 21 19 18 20 13 16 38]\n"
     ]
    }
   ],
   "source": [
    "for col in df.columns:\n",
    "    print(col,df[col].unique())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is mainly because of 3 columns, diag_1, diag_2 and diag_3 with 848, 923 and 954 unique values respectively. A lot of these values can be grouped into the same classes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "diag_cols = ['diag_1','diag_2','diag_3']\n",
    "for col in diag_cols:\n",
    "    df[col] = df[col].str.replace('E','-')\n",
    "    df[col] = df[col].str.replace('V','-')\n",
    "    condition = df[col].str.contains('250', na=False)\n",
    "    df.loc[condition,col] = '250'\n",
    "\n",
    "df[diag_cols] = df[diag_cols].astype(float)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "diag_1 ['Others' 'Neoplasms' 'Circulatory' 'Diabetes' 'Respiratory' 'Injury'\n",
      " 'External' 'Genitourinary' 'Muscoloskeletal' 'Digestive']\n",
      "diag_2 ['Diabetes' 'Neoplasms' 'Circulatory' 'Respiratory' 'Others' 'Injury'\n",
      " 'Muscoloskeletal' 'Genitourinary' 'Digestive' 'External']\n",
      "diag_3 ['Others' 'External' 'Circulatory' 'Diabetes' 'Respiratory' 'Injury'\n",
      " 'Neoplasms' 'Genitourinary' 'Digestive' 'Muscoloskeletal']\n"
     ]
    }
   ],
   "source": [
    "# diagnosis grouping based on ICD9 codes given in the paper\n",
    "for col in diag_cols:\n",
    "    df['temp']='Others'\n",
    "    df.loc[df[col]==250,'temp']='Diabetes'\n",
    "    df.loc[(df[col]>=390) & (df[col]<=458) | (df[col]==785),'temp']='Circulatory'\n",
    "    df.loc[(df[col]>=460) & (df[col]<=519) | (df[col]==786),'temp']='Respiratory' \n",
    "    df.loc[(df[col]>=520) & (df[col]<=579) | (df[col]==787),'temp']='Digestive'\n",
    "    df.loc[(df[col]>=580) & (df[col]<=629) | (df[col]==788),'temp']='Genitourinary'\n",
    "    df.loc[(df[col]>=800) & (df[col]<=999),'temp']='Injury'\n",
    "    df.loc[(df[col]>=710) & (df[col]<=739),'temp']='Muscoloskeletal'\n",
    "    df.loc[(df[col]>=140) & (df[col]<=239),'temp']='Neoplasms'\n",
    "    df.loc[df[col]<0,'temp']='External'\n",
    "    df[col]=df['temp']\n",
    "    df.drop('temp',axis=1,inplace=True)\n",
    "for col in diag_cols:\n",
    "    print(col,df[col].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Circulatory        20641\n",
      "Others             11210\n",
      "Respiratory         9544\n",
      "Digestive           6382\n",
      "Diabetes            5077\n",
      "Injury              4765\n",
      "Muscoloskeletal     3707\n",
      "Genitourinary       3520\n",
      "Neoplasms           2624\n",
      "External            1159\n",
      "Name: diag_1, dtype: int64\n",
      "----------------------\n",
      "Circulatory        21655\n",
      "Others             16084\n",
      "Diabetes            8457\n",
      "Respiratory         7306\n",
      "Genitourinary       5629\n",
      "Digestive           2839\n",
      "External            1833\n",
      "Neoplasms           1783\n",
      "Injury              1773\n",
      "Muscoloskeletal     1270\n",
      "Name: diag_2, dtype: int64\n",
      "----------------------\n",
      "Circulatory        20774\n",
      "Others             16658\n",
      "Diabetes           11892\n",
      "Respiratory         4810\n",
      "Genitourinary       4260\n",
      "External            3518\n",
      "Digestive           2620\n",
      "Injury              1404\n",
      "Muscoloskeletal     1352\n",
      "Neoplasms           1341\n",
      "Name: diag_3, dtype: int64\n",
      "----------------------\n"
     ]
    }
   ],
   "source": [
    "for col in diag_cols:\n",
    "    print(df[col].value_counts())\n",
    "    print(\"----------------------\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>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",
       "      <th>total_admissions</th>\n",
       "      <th>max_glu_serum_abnormal</th>\n",
       "      <th>A1CresultAbnormal</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Female</td>\n",
       "      <td>10</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>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>&gt;30</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>AfricanAmerican</td>\n",
       "      <td>Female</td>\n",
       "      <td>20</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>0</td>\n",
       "      <td>1</td>\n",
       "      <td>NO</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</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>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>NO</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</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>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>NO</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>50</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>31</td>\n",
       "      <td>6</td>\n",
       "      <td>16</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>&gt;30</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 46 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              race  gender  age  admission_type_id  discharge_disposition_id  \\\n",
       "1        Caucasian  Female   10                  1                         1   \n",
       "2  AfricanAmerican  Female   20                  1                         1   \n",
       "3        Caucasian    Male   30                  1                         1   \n",
       "4        Caucasian    Male   40                  1                         1   \n",
       "5        Caucasian    Male   50                  2                         1   \n",
       "\n",
       "   admission_source_id  time_in_hospital  num_lab_procedures  num_procedures  \\\n",
       "1                    7                 3                  59               0   \n",
       "2                    7                 2                  11               5   \n",
       "3                    7                 2                  44               1   \n",
       "4                    7                 1                  51               0   \n",
       "5                    2                 3                  31               6   \n",
       "\n",
       "   num_medications  ...  glipizide-metformin  glimepiride-pioglitazone  \\\n",
       "1               18  ...                   No                        No   \n",
       "2               13  ...                   No                        No   \n",
       "3               16  ...                   No                        No   \n",
       "4                8  ...                   No                        No   \n",
       "5               16  ...                   No                        No   \n",
       "\n",
       "   metformin-rosiglitazone metformin-pioglitazone change diabetesMed  \\\n",
       "1                       No                     No      1           1   \n",
       "2                       No                     No      0           1   \n",
       "3                       No                     No      1           1   \n",
       "4                       No                     No      1           1   \n",
       "5                       No                     No      0           1   \n",
       "\n",
       "   readmitted total_admissions max_glu_serum_abnormal A1CresultAbnormal  \n",
       "1         >30                0                   -100              -100  \n",
       "2          NO                0                   -100              -100  \n",
       "3          NO                0                   -100              -100  \n",
       "4          NO                0                   -100              -100  \n",
       "5         >30                0                   -100              -100  \n",
       "\n",
       "[5 rows x 46 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['age'] = df['age'].str[1:].str.split('-',expand=True)[0]\n",
    "df['age'] = df['age'].astype(int)\n",
    "max_glu_serum_dict = {'None':-100,\n",
    "                      'Norm':0,\n",
    "                      '>200':1,\n",
    "                      '>300':1\n",
    "                     }\n",
    "df['max_glu_serum_abnormal'] = df['max_glu_serum'].replace(max_glu_serum_dict)\n",
    "df.drop('max_glu_serum',axis=1,inplace=True)\n",
    "\n",
    "A1Cresult_dict = {'None':-100,\n",
    "                  'Norm':0,\n",
    "                  '>7':1,\n",
    "                  '>8':1\n",
    "                 }\n",
    "df['A1CresultAbnormal'] = df['A1Cresult'].replace(A1Cresult_dict)\n",
    "df.drop('A1Cresult',axis=1,inplace=True)\n",
    "\n",
    "change_dict = {'No':0,\n",
    "               'Ch':1\n",
    "              }\n",
    "df['change'] = df['change'].replace(change_dict)\n",
    "\n",
    "diabetesMed_dict = {'No':0,\n",
    "                    'Yes':1\n",
    "                   }\n",
    "df['diabetesMed'] = df['diabetesMed'].replace(diabetesMed_dict)\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>metformin-rosiglitazone</th>\n",
       "      <th>metformin-pioglitazone</th>\n",
       "      <th>change</th>\n",
       "      <th>diabetesMed</th>\n",
       "      <th>readmitted</th>\n",
       "      <th>total_admissions</th>\n",
       "      <th>max_glu_serum_abnormal</th>\n",
       "      <th>A1CresultAbnormal</th>\n",
       "      <th>medication_changes</th>\n",
       "      <th>within30</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Female</td>\n",
       "      <td>10</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>-200</td>\n",
       "      <td>-200</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>AfricanAmerican</td>\n",
       "      <td>Female</td>\n",
       "      <td>20</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>-200</td>\n",
       "      <td>-200</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</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>-200</td>\n",
       "      <td>-200</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</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>-200</td>\n",
       "      <td>-200</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>Caucasian</td>\n",
       "      <td>Male</td>\n",
       "      <td>50</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>31</td>\n",
       "      <td>6</td>\n",
       "      <td>16</td>\n",
       "      <td>...</td>\n",
       "      <td>-200</td>\n",
       "      <td>-200</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-100</td>\n",
       "      <td>-100</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 48 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              race  gender  age  admission_type_id  discharge_disposition_id  \\\n",
       "1        Caucasian  Female   10                  1                         1   \n",
       "2  AfricanAmerican  Female   20                  1                         1   \n",
       "3        Caucasian    Male   30                  1                         1   \n",
       "4        Caucasian    Male   40                  1                         1   \n",
       "5        Caucasian    Male   50                  2                         1   \n",
       "\n",
       "   admission_source_id  time_in_hospital  num_lab_procedures  num_procedures  \\\n",
       "1                    7                 3                  59               0   \n",
       "2                    7                 2                  11               5   \n",
       "3                    7                 2                  44               1   \n",
       "4                    7                 1                  51               0   \n",
       "5                    2                 3                  31               6   \n",
       "\n",
       "   num_medications  ...  metformin-rosiglitazone  metformin-pioglitazone  \\\n",
       "1               18  ...                     -200                    -200   \n",
       "2               13  ...                     -200                    -200   \n",
       "3               16  ...                     -200                    -200   \n",
       "4                8  ...                     -200                    -200   \n",
       "5               16  ...                     -200                    -200   \n",
       "\n",
       "   change diabetesMed readmitted total_admissions  max_glu_serum_abnormal  \\\n",
       "1       1           1          1                0                    -100   \n",
       "2       0           1          0                0                    -100   \n",
       "3       1           1          0                0                    -100   \n",
       "4       1           1          0                0                    -100   \n",
       "5       0           1          1                0                    -100   \n",
       "\n",
       "   A1CresultAbnormal  medication_changes  within30  \n",
       "1               -100                   1         0  \n",
       "2               -100                   0         0  \n",
       "3               -100                   1         0  \n",
       "4               -100                   0         0  \n",
       "5               -100                   0         0  \n",
       "\n",
       "[5 rows x 48 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d24_feature_dict = {'Up':10,\n",
    "                    'Down':-10,\n",
    "                    'Steady':0,\n",
    "                    'No':-200\n",
    "                   }\n",
    "d24_cols = ['metformin','repaglinide','nateglinide','chlorpropamide',\n",
    " 'glimepiride','acetohexamide','glipizide','glyburide',\n",
    " 'tolbutamide','pioglitazone','rosiglitazone','acarbose',\n",
    " 'miglitol','troglitazone','tolazamide','examide',\n",
    " 'citoglipton','insulin','glyburide-metformin','glipizide-metformin',\n",
    " 'glimepiride-pioglitazone','metformin-rosiglitazone','metformin-pioglitazone']\n",
    "\n",
    "for col in d24_cols:\n",
    "    colname = str(col) + 'temp'\n",
    "    df[colname] = df[col].apply(lambda x: 0 if (x == 'No' or x == 'Steady') else 1)\n",
    "df['medication_changes'] = 0\n",
    "for col in d24_cols:\n",
    "    colname = str(col) + 'temp'\n",
    "    df['medication_changes'] = df['medication_changes'] + df[colname]\n",
    "    del df[colname]\n",
    "\n",
    "for col in d24_cols:\n",
    "    df[col] = df[col].replace(d24_feature_dict)\n",
    "\n",
    "df['within30'] = np.where(df['readmitted']=='<30',1,0)\n",
    "df['readmitted'] = np.where(df['readmitted']=='NO',0,1)\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1     41323\n",
       "3      8987\n",
       "6      8253\n",
       "18     2348\n",
       "2      1401\n",
       "22     1267\n",
       "5       794\n",
       "25      536\n",
       "4       506\n",
       "7       388\n",
       "13      357\n",
       "14      346\n",
       "23      272\n",
       "28       77\n",
       "8        69\n",
       "15       30\n",
       "24       29\n",
       "9        12\n",
       "17       11\n",
       "10        6\n",
       "16        5\n",
       "27        4\n",
       "12        2\n",
       "Name: discharge_disposition_id, dtype: int64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 11,19,20,21 represent the patient expired at hospital or at home. We should remove these from our dataset as they cannot be \n",
    "# readmitted.\n",
    "df = df[~df['discharge_disposition_id'].isin([11,19,20,21])]\n",
    "df['discharge_disposition_id'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "race                        0\n",
       "gender                      0\n",
       "age                         0\n",
       "admission_type_id           0\n",
       "discharge_disposition_id    0\n",
       "admission_source_id         0\n",
       "time_in_hospital            0\n",
       "num_lab_procedures          0\n",
       "num_procedures              0\n",
       "num_medications             0\n",
       "number_outpatient           0\n",
       "number_emergency            0\n",
       "number_inpatient            0\n",
       "diag_1                      0\n",
       "diag_2                      0\n",
       "diag_3                      0\n",
       "number_diagnoses            0\n",
       "metformin                   0\n",
       "repaglinide                 0\n",
       "nateglinide                 0\n",
       "chlorpropamide              0\n",
       "glimepiride                 0\n",
       "acetohexamide               0\n",
       "glipizide                   0\n",
       "glyburide                   0\n",
       "tolbutamide                 0\n",
       "pioglitazone                0\n",
       "rosiglitazone               0\n",
       "acarbose                    0\n",
       "miglitol                    0\n",
       "troglitazone                0\n",
       "tolazamide                  0\n",
       "examide                     0\n",
       "citoglipton                 0\n",
       "insulin                     0\n",
       "glyburide-metformin         0\n",
       "glipizide-metformin         0\n",
       "glimepiride-pioglitazone    0\n",
       "metformin-rosiglitazone     0\n",
       "metformin-pioglitazone      0\n",
       "change                      0\n",
       "diabetesMed                 0\n",
       "readmitted                  0\n",
       "total_admissions            0\n",
       "max_glu_serum_abnormal      0\n",
       "A1CresultAbnormal           0\n",
       "medication_changes          0\n",
       "within30                    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Recode all the 'Not Available/Null' categories into one \n",
    "df['admission_type_id'] = df['admission_type_id'].replace({6: 5, 8: 5})\n",
    "df['admission_source_id'] = df['admission_source_id'].replace({15: 9, 17: 9, 20: 9, 21: 9})\n",
    "df['discharge_disposition_id'] = df['discharge_disposition_id'].replace({25: 18, 26: 18})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>time_in_hospital</th>\n",
       "      <th>num_lab_procedures</th>\n",
       "      <th>num_procedures</th>\n",
       "      <th>num_medications</th>\n",
       "      <th>number_outpatient</th>\n",
       "      <th>number_emergency</th>\n",
       "      <th>number_inpatient</th>\n",
       "      <th>number_diagnoses</th>\n",
       "      <th>metformin</th>\n",
       "      <th>...</th>\n",
       "      <th>admission_source_id_6</th>\n",
       "      <th>admission_source_id_7</th>\n",
       "      <th>admission_source_id_8</th>\n",
       "      <th>admission_source_id_9</th>\n",
       "      <th>admission_source_id_10</th>\n",
       "      <th>admission_source_id_11</th>\n",
       "      <th>admission_source_id_13</th>\n",
       "      <th>admission_source_id_14</th>\n",
       "      <th>admission_source_id_22</th>\n",
       "      <th>admission_source_id_25</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>59</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>20</td>\n",
       "      <td>2</td>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>30</td>\n",
       "      <td>2</td>\n",
       "      <td>44</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>40</td>\n",
       "      <td>1</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>50</td>\n",
       "      <td>3</td>\n",
       "      <td>31</td>\n",
       "      <td>6</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 120 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  time_in_hospital  num_lab_procedures  num_procedures  num_medications  \\\n",
       "1   10                 3                  59               0               18   \n",
       "2   20                 2                  11               5               13   \n",
       "3   30                 2                  44               1               16   \n",
       "4   40                 1                  51               0                8   \n",
       "5   50                 3                  31               6               16   \n",
       "\n",
       "   number_outpatient  number_emergency  number_inpatient  number_diagnoses  \\\n",
       "1                  0                 0                 0                 9   \n",
       "2                  2                 0                 1                 6   \n",
       "3                  0                 0                 0                 7   \n",
       "4                  0                 0                 0                 5   \n",
       "5                  0                 0                 0                 9   \n",
       "\n",
       "   metformin  ...  admission_source_id_6  admission_source_id_7  \\\n",
       "1       -200  ...                      0                      1   \n",
       "2       -200  ...                      0                      1   \n",
       "3       -200  ...                      0                      1   \n",
       "4       -200  ...                      0                      1   \n",
       "5       -200  ...                      0                      0   \n",
       "\n",
       "   admission_source_id_8  admission_source_id_9  admission_source_id_10  \\\n",
       "1                      0                      0                       0   \n",
       "2                      0                      0                       0   \n",
       "3                      0                      0                       0   \n",
       "4                      0                      0                       0   \n",
       "5                      0                      0                       0   \n",
       "\n",
       "   admission_source_id_11  admission_source_id_13  admission_source_id_14  \\\n",
       "1                       0                       0                       0   \n",
       "2                       0                       0                       0   \n",
       "3                       0                       0                       0   \n",
       "4                       0                       0                       0   \n",
       "5                       0                       0                       0   \n",
       "\n",
       "   admission_source_id_22  admission_source_id_25  \n",
       "1                       0                       0  \n",
       "2                       0                       0  \n",
       "3                       0                       0  \n",
       "4                       0                       0  \n",
       "5                       0                       0  \n",
       "\n",
       "[5 rows x 120 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cat_data = [\"race\",\"gender\",\"diag_1\",\"diag_2\",\"diag_3\",\n",
    " \"admission_type_id\",\"discharge_disposition_id\",\"admission_source_id\"]\n",
    "for col in cat_data:\n",
    "    df = pd.concat([df.drop(col, axis=1), pd.get_dummies(df[col],prefix=col)], axis=1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>time_in_hospital</th>\n",
       "      <th>num_lab_procedures</th>\n",
       "      <th>num_procedures</th>\n",
       "      <th>num_medications</th>\n",
       "      <th>number_outpatient</th>\n",
       "      <th>number_emergency</th>\n",
       "      <th>number_inpatient</th>\n",
       "      <th>number_diagnoses</th>\n",
       "      <th>metformin</th>\n",
       "      <th>...</th>\n",
       "      <th>admission_source_id_6</th>\n",
       "      <th>admission_source_id_7</th>\n",
       "      <th>admission_source_id_8</th>\n",
       "      <th>admission_source_id_9</th>\n",
       "      <th>admission_source_id_10</th>\n",
       "      <th>admission_source_id_11</th>\n",
       "      <th>admission_source_id_13</th>\n",
       "      <th>admission_source_id_14</th>\n",
       "      <th>admission_source_id_22</th>\n",
       "      <th>admission_source_id_25</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>59</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>2</td>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>30</td>\n",
       "      <td>2</td>\n",
       "      <td>44</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>40</td>\n",
       "      <td>1</td>\n",
       "      <td>51</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>50</td>\n",
       "      <td>3</td>\n",
       "      <td>31</td>\n",
       "      <td>6</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>-200</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 120 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  time_in_hospital  num_lab_procedures  num_procedures  num_medications  \\\n",
       "0   10                 3                  59               0               18   \n",
       "1   20                 2                  11               5               13   \n",
       "2   30                 2                  44               1               16   \n",
       "3   40                 1                  51               0                8   \n",
       "4   50                 3                  31               6               16   \n",
       "\n",
       "   number_outpatient  number_emergency  number_inpatient  number_diagnoses  \\\n",
       "0                  0                 0                 0                 9   \n",
       "1                  2                 0                 1                 6   \n",
       "2                  0                 0                 0                 7   \n",
       "3                  0                 0                 0                 5   \n",
       "4                  0                 0                 0                 9   \n",
       "\n",
       "   metformin  ...  admission_source_id_6  admission_source_id_7  \\\n",
       "0       -200  ...                      0                      1   \n",
       "1       -200  ...                      0                      1   \n",
       "2       -200  ...                      0                      1   \n",
       "3       -200  ...                      0                      1   \n",
       "4       -200  ...                      0                      0   \n",
       "\n",
       "   admission_source_id_8  admission_source_id_9  admission_source_id_10  \\\n",
       "0                      0                      0                       0   \n",
       "1                      0                      0                       0   \n",
       "2                      0                      0                       0   \n",
       "3                      0                      0                       0   \n",
       "4                      0                      0                       0   \n",
       "\n",
       "   admission_source_id_11  admission_source_id_13  admission_source_id_14  \\\n",
       "0                       0                       0                       0   \n",
       "1                       0                       0                       0   \n",
       "2                       0                       0                       0   \n",
       "3                       0                       0                       0   \n",
       "4                       0                       0                       0   \n",
       "\n",
       "   admission_source_id_22  admission_source_id_25  \n",
       "0                       0                       0  \n",
       "1                       0                       0  \n",
       "2                       0                       0  \n",
       "3                       0                       0  \n",
       "4                       0                       0  \n",
       "\n",
       "[5 rows x 120 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.reset_index(inplace=True,drop=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    50258\n",
       "1    16765\n",
       "Name: readmitted, dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"readmitted\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    63859\n",
       "1     3164\n",
       "Name: within30, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['within30'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(67023, 120)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv(\"Clean_Data_LogReg.csv\",index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}