30239_ PopMarcus.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A2: prezicere supravietuire pacient COVID\n",
    "Marcus Pop <br>\n",
    "Group 30239"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import re\n",
    "import datetime"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Obs: S-a lucrat pe un dataset mai vechi. Acesta este atasat in arhiva. Impartirea pe train/test se face programatic."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a href=\"https://datarepository.wolframcloud.com/resources/Patient-Medical-Data-for-Novel-Coronavirus-COVID-19\">Dataset link.</a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "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>Sex</th>\n",
       "      <th>City</th>\n",
       "      <th>AdministrativeDivision</th>\n",
       "      <th>Country</th>\n",
       "      <th>GeoPosition</th>\n",
       "      <th>DateOfOnsetSymptoms</th>\n",
       "      <th>DateOfAdmissionHospital</th>\n",
       "      <th>DateOfConfirmation</th>\n",
       "      <th>Symptoms</th>\n",
       "      <th>...</th>\n",
       "      <th>TravelHistoryLocation</th>\n",
       "      <th>ReportedMarketExposure</th>\n",
       "      <th>ReportedMarketExposureComment</th>\n",
       "      <th>ChronicDiseaseQ</th>\n",
       "      <th>ChronicDiseases</th>\n",
       "      <th>SequenceAvailable</th>\n",
       "      <th>DischargedQ</th>\n",
       "      <th>DeathQ</th>\n",
       "      <th>DateOfDeath</th>\n",
       "      <th>DateOfDischarge</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>48</td>\n",
       "      <td>Entity[\"Gender\", \"Female\"]</td>\n",
       "      <td>Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]</td>\n",
       "      <td>Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...</td>\n",
       "      <td>Entity[\"Country\", \"China\"]</td>\n",
       "      <td>GeoPosition[{30.625059999999998, 114.3421}]</td>\n",
       "      <td>DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\", \"fatigue\", \"fever\", \"sputum\", \"myalg...</td>\n",
       "      <td>...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>{\"diabetes\", \"cerebral infarction\"}</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>DateObject[{2020, 1, 20}, \"Day\", \"Gregorian\", ...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Interval[{15, 88}]</td>\n",
       "      <td>Entity[\"Gender\", \"Female\"]</td>\n",
       "      <td>Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]</td>\n",
       "      <td>Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...</td>\n",
       "      <td>Entity[\"Country\", \"China\"]</td>\n",
       "      <td>GeoPosition[{30.625059999999998, 114.3421}]</td>\n",
       "      <td>DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"shortness of breath\", \"respiratory symptoms\"}</td>\n",
       "      <td>...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>61</td>\n",
       "      <td>Entity[\"Gender\", \"Male\"]</td>\n",
       "      <td>Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]</td>\n",
       "      <td>Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...</td>\n",
       "      <td>Entity[\"Country\", \"China\"]</td>\n",
       "      <td>GeoPosition[{30.625059999999998, 114.3421}]</td>\n",
       "      <td>DateObject[{2019, 12, 20}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 27}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>DateObject[{2020, 1, 9}, \"Day\", \"Gregorian\", -5.]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>58</td>\n",
       "      <td>Entity[\"Gender\", \"Male\"]</td>\n",
       "      <td>Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]</td>\n",
       "      <td>Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...</td>\n",
       "      <td>Entity[\"Country\", \"China\"]</td>\n",
       "      <td>GeoPosition[{30.625059999999998, 114.3421}]</td>\n",
       "      <td>DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2020, 1, 1}, \"Day\", \"Gregorian\", -5.]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\", \"sputum\", \"fever\"}</td>\n",
       "      <td>...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>DateObject[{2020, 1, 23}, \"Day\", \"Gregorian\", ...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>66</td>\n",
       "      <td>Entity[\"Gender\", \"Male\"]</td>\n",
       "      <td>Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]</td>\n",
       "      <td>Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...</td>\n",
       "      <td>Entity[\"Country\", \"China\"]</td>\n",
       "      <td>GeoPosition[{30.625059999999998, 114.3421}]</td>\n",
       "      <td>DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\"}</td>\n",
       "      <td>...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>DateObject[{2020, 1, 21}, \"Day\", \"Gregorian\", ...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Age                         Sex  \\\n",
       "0                  48  Entity[\"Gender\", \"Female\"]   \n",
       "1  Interval[{15, 88}]  Entity[\"Gender\", \"Female\"]   \n",
       "2                  61    Entity[\"Gender\", \"Male\"]   \n",
       "3                  58    Entity[\"Gender\", \"Male\"]   \n",
       "4                  66    Entity[\"Gender\", \"Male\"]   \n",
       "\n",
       "                                          City  \\\n",
       "0  Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]   \n",
       "1  Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]   \n",
       "2  Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]   \n",
       "3  Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]   \n",
       "4  Entity[\"City\", {\"Wuhan\", \"Hubei\", \"China\"}]   \n",
       "\n",
       "                              AdministrativeDivision  \\\n",
       "0  Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...   \n",
       "1  Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...   \n",
       "2  Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...   \n",
       "3  Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...   \n",
       "4  Entity[\"AdministrativeDivision\", {\"Hubei\", \"Ch...   \n",
       "\n",
       "                      Country                                  GeoPosition  \\\n",
       "0  Entity[\"Country\", \"China\"]  GeoPosition[{30.625059999999998, 114.3421}]   \n",
       "1  Entity[\"Country\", \"China\"]  GeoPosition[{30.625059999999998, 114.3421}]   \n",
       "2  Entity[\"Country\", \"China\"]  GeoPosition[{30.625059999999998, 114.3421}]   \n",
       "3  Entity[\"Country\", \"China\"]  GeoPosition[{30.625059999999998, 114.3421}]   \n",
       "4  Entity[\"Country\", \"China\"]  GeoPosition[{30.625059999999998, 114.3421}]   \n",
       "\n",
       "                                 DateOfOnsetSymptoms  \\\n",
       "0  DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...   \n",
       "1  DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...   \n",
       "2  DateObject[{2019, 12, 20}, \"Day\", \"Gregorian\",...   \n",
       "3  DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...   \n",
       "4  DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...   \n",
       "\n",
       "                             DateOfAdmissionHospital       DateOfConfirmation  \\\n",
       "0  DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "1  DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "2  DateObject[{2019, 12, 27}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "3  DateObject[{2020, 1, 1}, \"Day\", \"Gregorian\", -5.]  Missing[\"NotAvailable\"]   \n",
       "4  DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "\n",
       "                                            Symptoms  ...  \\\n",
       "0  {\"cough\", \"fatigue\", \"fever\", \"sputum\", \"myalg...  ...   \n",
       "1    {\"shortness of breath\", \"respiratory symptoms\"}  ...   \n",
       "2                            Missing[\"NotAvailable\"]  ...   \n",
       "3                       {\"cough\", \"sputum\", \"fever\"}  ...   \n",
       "4                                          {\"cough\"}  ...   \n",
       "\n",
       "     TravelHistoryLocation   ReportedMarketExposure  \\\n",
       "0  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "1  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "3  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "4  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "\n",
       "  ReportedMarketExposureComment ChronicDiseaseQ  \\\n",
       "0       Missing[\"NotAvailable\"]            True   \n",
       "1       Missing[\"NotAvailable\"]           False   \n",
       "2       Missing[\"NotAvailable\"]           False   \n",
       "3       Missing[\"NotAvailable\"]           False   \n",
       "4       Missing[\"NotAvailable\"]           False   \n",
       "\n",
       "                       ChronicDiseases        SequenceAvailable  \\\n",
       "0  {\"diabetes\", \"cerebral infarction\"}  Missing[\"NotAvailable\"]   \n",
       "1              Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2              Missing[\"NotAvailable\"]                     True   \n",
       "3              Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "4              Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "\n",
       "               DischargedQ DeathQ  \\\n",
       "0  Missing[\"NotAvailable\"]   True   \n",
       "1  Missing[\"NotAvailable\"]   True   \n",
       "2  Missing[\"NotAvailable\"]   True   \n",
       "3  Missing[\"NotAvailable\"]   True   \n",
       "4  Missing[\"NotAvailable\"]   True   \n",
       "\n",
       "                                         DateOfDeath          DateOfDischarge  \n",
       "0  DateObject[{2020, 1, 20}, \"Day\", \"Gregorian\", ...  Missing[\"NotAvailable\"]  \n",
       "1                            Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  \n",
       "2  DateObject[{2020, 1, 9}, \"Day\", \"Gregorian\", -5.]  Missing[\"NotAvailable\"]  \n",
       "3  DateObject[{2020, 1, 23}, \"Day\", \"Gregorian\", ...  Missing[\"NotAvailable\"]  \n",
       "4  DateObject[{2020, 1, 21}, \"Day\", \"Gregorian\", ...  Missing[\"NotAvailable\"]  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "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>Sex</th>\n",
       "      <th>City</th>\n",
       "      <th>AdministrativeDivision</th>\n",
       "      <th>Country</th>\n",
       "      <th>GeoPosition</th>\n",
       "      <th>DateOfOnsetSymptoms</th>\n",
       "      <th>DateOfAdmissionHospital</th>\n",
       "      <th>DateOfConfirmation</th>\n",
       "      <th>Symptoms</th>\n",
       "      <th>...</th>\n",
       "      <th>TravelHistoryLocation</th>\n",
       "      <th>ReportedMarketExposure</th>\n",
       "      <th>ReportedMarketExposureComment</th>\n",
       "      <th>ChronicDiseaseQ</th>\n",
       "      <th>ChronicDiseases</th>\n",
       "      <th>SequenceAvailable</th>\n",
       "      <th>DischargedQ</th>\n",
       "      <th>DeathQ</th>\n",
       "      <th>DateOfDeath</th>\n",
       "      <th>DateOfDischarge</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>...</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "      <td>19467</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>209</td>\n",
       "      <td>3</td>\n",
       "      <td>720</td>\n",
       "      <td>463</td>\n",
       "      <td>137</td>\n",
       "      <td>1834</td>\n",
       "      <td>103</td>\n",
       "      <td>83</td>\n",
       "      <td>123</td>\n",
       "      <td>319</td>\n",
       "      <td>...</td>\n",
       "      <td>587</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>67</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>40</td>\n",
       "      <td>66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Entity[\"Gender\", \"Male\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Entity[\"Country\", \"UnitedStates\"]</td>\n",
       "      <td>GeoPosition[{25.6135986, -80.565963}]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>DateObject[{2020, 3, 26}, \"Day\", \"Gregorian\", ...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>2083</td>\n",
       "      <td>9717</td>\n",
       "      <td>12178</td>\n",
       "      <td>4226</td>\n",
       "      <td>6002</td>\n",
       "      <td>1079</td>\n",
       "      <td>15119</td>\n",
       "      <td>17601</td>\n",
       "      <td>1380</td>\n",
       "      <td>17781</td>\n",
       "      <td>...</td>\n",
       "      <td>14355</td>\n",
       "      <td>19392</td>\n",
       "      <td>19454</td>\n",
       "      <td>18717</td>\n",
       "      <td>19318</td>\n",
       "      <td>19461</td>\n",
       "      <td>19370</td>\n",
       "      <td>19317</td>\n",
       "      <td>19321</td>\n",
       "      <td>19106</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                            Age                       Sex  \\\n",
       "count                     19467                     19467   \n",
       "unique                      209                         3   \n",
       "top     Missing[\"NotAvailable\"]  Entity[\"Gender\", \"Male\"]   \n",
       "freq                       2083                      9717   \n",
       "\n",
       "                           City   AdministrativeDivision  \\\n",
       "count                     19467                    19467   \n",
       "unique                      720                      463   \n",
       "top     Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "freq                      12178                     4226   \n",
       "\n",
       "                                  Country  \\\n",
       "count                               19467   \n",
       "unique                                137   \n",
       "top     Entity[\"Country\", \"UnitedStates\"]   \n",
       "freq                                 6002   \n",
       "\n",
       "                                  GeoPosition      DateOfOnsetSymptoms  \\\n",
       "count                                   19467                    19467   \n",
       "unique                                   1834                      103   \n",
       "top     GeoPosition[{25.6135986, -80.565963}]  Missing[\"NotAvailable\"]   \n",
       "freq                                     1079                    15119   \n",
       "\n",
       "        DateOfAdmissionHospital  \\\n",
       "count                     19467   \n",
       "unique                       83   \n",
       "top     Missing[\"NotAvailable\"]   \n",
       "freq                      17601   \n",
       "\n",
       "                                       DateOfConfirmation  \\\n",
       "count                                               19467   \n",
       "unique                                                123   \n",
       "top     DateObject[{2020, 3, 26}, \"Day\", \"Gregorian\", ...   \n",
       "freq                                                 1380   \n",
       "\n",
       "                       Symptoms  ...    TravelHistoryLocation  \\\n",
       "count                     19467  ...                    19467   \n",
       "unique                      319  ...                      587   \n",
       "top     Missing[\"NotAvailable\"]  ...  Missing[\"NotAvailable\"]   \n",
       "freq                      17781  ...                    14355   \n",
       "\n",
       "         ReportedMarketExposure ReportedMarketExposureComment ChronicDiseaseQ  \\\n",
       "count                     19467                         19467           19467   \n",
       "unique                        3                             7               2   \n",
       "top     Missing[\"NotAvailable\"]       Missing[\"NotAvailable\"]           False   \n",
       "freq                      19392                         19454           18717   \n",
       "\n",
       "                ChronicDiseases        SequenceAvailable  \\\n",
       "count                     19467                    19467   \n",
       "unique                       67                        2   \n",
       "top     Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "freq                      19318                    19461   \n",
       "\n",
       "                    DischargedQ                   DeathQ  \\\n",
       "count                     19467                    19467   \n",
       "unique                        2                        2   \n",
       "top     Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "freq                      19370                    19317   \n",
       "\n",
       "                    DateOfDeath          DateOfDischarge  \n",
       "count                     19467                    19467  \n",
       "unique                       40                       66  \n",
       "top     Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  \n",
       "freq                      19321                    19106  \n",
       "\n",
       "[4 rows x 23 columns]"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### As we can we, the dataset is rather imbalanced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Missing[\"NotAvailable\"]    19317\n",
       "True                         150\n",
       "Name: DeathQ, dtype: int64"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['DeathQ'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One hot encoding for travel history\n",
    "\n",
    "Travel history location data is represented as a string of entities depicting the path the pacient has taken through different map points.\n",
    "\n",
    "For a better representation we are going to represent it using one hot encoding using the unique locations as bits for the following reasons:\n",
    "- we are not interested in order, just simply the locations\n",
    "- it is easier to work and represent data in this way"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "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>Entity[\"AdministrativeDivision\", {\"AbuDhabi\", \"UnitedArabEmirates\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Accra\", \"GreaterAccra\", \"Ghana\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"AddisAbaba\", \"Ethiopia\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Aichi\", \"Japan\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Alabama\", \"UnitedStates\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Alberta\", \"Canada\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Alsace\", \"France\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Amsterdam\", \"NoordHolland\", \"Netherlands\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Andalo\", \"Trento\", \"TrentinoAltoAdige\", \"Italy\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Anhui\", \"China\"}]</th>\n",
       "      <th>...</th>\n",
       "      <th>Entity[\"Country\", \"UnitedStates\"]</th>\n",
       "      <th>Entity[\"Country\", \"Uruguay\"]</th>\n",
       "      <th>Entity[\"Country\", \"VaticanCity\"]</th>\n",
       "      <th>Entity[\"Country\", \"Venezuela\"]</th>\n",
       "      <th>Entity[\"Country\", \"Vietnam\"]</th>\n",
       "      <th>Entity[\"Country\", \"Zambia\"]</th>\n",
       "      <th>Entity[\"Country\", \"Zimbabwe\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Africa\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Asia\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Europe\"]</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 362 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Entity[\"AdministrativeDivision\", {\"AbuDhabi\", \"UnitedArabEmirates\"}]  \\\n",
       "0                                              False                      \n",
       "1                                              False                      \n",
       "2                                              False                      \n",
       "3                                              False                      \n",
       "4                                              False                      \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Accra\", \"GreaterAccra\", \"Ghana\"}]  \\\n",
       "0                                              False                      \n",
       "1                                              False                      \n",
       "2                                              False                      \n",
       "3                                              False                      \n",
       "4                                              False                      \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"AddisAbaba\", \"Ethiopia\"}]  \\\n",
       "0                                              False              \n",
       "1                                              False              \n",
       "2                                              False              \n",
       "3                                              False              \n",
       "4                                              False              \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Aichi\", \"Japan\"}]  \\\n",
       "0                                              False      \n",
       "1                                              False      \n",
       "2                                              False      \n",
       "3                                              False      \n",
       "4                                              False      \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Alabama\", \"UnitedStates\"}]  \\\n",
       "0                                              False               \n",
       "1                                              False               \n",
       "2                                              False               \n",
       "3                                              False               \n",
       "4                                              False               \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Alberta\", \"Canada\"}]  \\\n",
       "0                                              False         \n",
       "1                                              False         \n",
       "2                                              False         \n",
       "3                                              False         \n",
       "4                                              False         \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Alsace\", \"France\"}]  \\\n",
       "0                                              False        \n",
       "1                                              False        \n",
       "2                                              False        \n",
       "3                                              False        \n",
       "4                                              False        \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Amsterdam\", \"NoordHolland\", \"Netherlands\"}]  \\\n",
       "0                                              False                                \n",
       "1                                              False                                \n",
       "2                                              False                                \n",
       "3                                              False                                \n",
       "4                                              False                                \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Andalo\", \"Trento\", \"TrentinoAltoAdige\", \"Italy\"}]  \\\n",
       "0                                              False                                      \n",
       "1                                              False                                      \n",
       "2                                              False                                      \n",
       "3                                              False                                      \n",
       "4                                              False                                      \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Anhui\", \"China\"}]  ...  \\\n",
       "0                                              False     ...   \n",
       "1                                              False     ...   \n",
       "2                                              False     ...   \n",
       "3                                              False     ...   \n",
       "4                                              False     ...   \n",
       "\n",
       "   Entity[\"Country\", \"UnitedStates\"]  Entity[\"Country\", \"Uruguay\"]  \\\n",
       "0                              False                         False   \n",
       "1                              False                         False   \n",
       "2                              False                         False   \n",
       "3                              False                         False   \n",
       "4                              False                         False   \n",
       "\n",
       "   Entity[\"Country\", \"VaticanCity\"]  Entity[\"Country\", \"Venezuela\"]  \\\n",
       "0                             False                           False   \n",
       "1                             False                           False   \n",
       "2                             False                           False   \n",
       "3                             False                           False   \n",
       "4                             False                           False   \n",
       "\n",
       "   Entity[\"Country\", \"Vietnam\"]  Entity[\"Country\", \"Zambia\"]  \\\n",
       "0                         False                        False   \n",
       "1                         False                        False   \n",
       "2                         False                        False   \n",
       "3                         False                        False   \n",
       "4                         False                        False   \n",
       "\n",
       "   Entity[\"Country\", \"Zimbabwe\"]  Entity[\"GeographicRegion\", \"Africa\"]  \\\n",
       "0                          False                                 False   \n",
       "1                          False                                 False   \n",
       "2                          False                                 False   \n",
       "3                          False                                 False   \n",
       "4                          False                                 False   \n",
       "\n",
       "   Entity[\"GeographicRegion\", \"Asia\"]  Entity[\"GeographicRegion\", \"Europe\"]  \n",
       "0                               False                                 False  \n",
       "1                               False                                 False  \n",
       "2                               False                                 False  \n",
       "3                               False                                 False  \n",
       "4                               False                                 False  \n",
       "\n",
       "[5 rows x 362 columns]"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# transform travel history from locations string to list of locations\n",
    "travel_col = 'TravelHistoryLocation'\n",
    "transform = lambda x: re.findall(r\"Entity\\[[^\\]\\[]+\\]\", x)\n",
    "travel_lists = df[travel_col].apply(transform)\n",
    "# get the unique locations found in the dataset\n",
    "locs = np.concatenate(travel_lists.values)\n",
    "unique_locs = np.unique(locs)\n",
    "# transform travel history from list to one hot encoding\n",
    "onehot = np.array([np.in1d(unique_locs, list_col) for list_col in travel_lists])\n",
    "# create new dataframe using the unique locations as column headers\n",
    "onehot_frame = pd.DataFrame(onehot, columns=unique_locs)\n",
    "onehot_frame.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop the old travel history column from the origin data frame\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "travel_dropped = df.drop(columns=['TravelHistoryLocation'], axis=1)\n",
    "df = pd.concat([travel_dropped, onehot_frame], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Splitting the geoposition entities into latitude and longitude, plus dropping the now redundant location features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "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>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>19467.000000</td>\n",
       "      <td>19467.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>22.315788</td>\n",
       "      <td>-12.287439</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>18.706135</td>\n",
       "      <td>93.692944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-43.570000</td>\n",
       "      <td>-159.727596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>15.369250</td>\n",
       "      <td>-82.308992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>25.613599</td>\n",
       "      <td>-77.015390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>33.727220</td>\n",
       "      <td>106.174300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>69.649468</td>\n",
       "      <td>174.740000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Latitude     Longitude\n",
       "count  19467.000000  19467.000000\n",
       "mean      22.315788    -12.287439\n",
       "std       18.706135     93.692944\n",
       "min      -43.570000   -159.727596\n",
       "25%       15.369250    -82.308992\n",
       "50%       25.613599    -77.015390\n",
       "75%       33.727220    106.174300\n",
       "max       69.649468    174.740000"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# transform the geoposition column in two seperate column(lat, long)\n",
    "transform = lambda x: [float(x) for x in re.findall(r\"[+-]?[0-9]*[.]?[0-9]+\", x)]\n",
    "latlong = df['GeoPosition'].apply(transform).values\n",
    "# from array of lists to ndarray\n",
    "latlong = np.vstack(latlong)\n",
    "latlong_frame=pd.DataFrame(latlong, columns=['Latitude', 'Longitude'])\n",
    "latlong_frame.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop redundant location columns and add the latitude/longitude frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "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>Sex</th>\n",
       "      <th>DateOfOnsetSymptoms</th>\n",
       "      <th>DateOfAdmissionHospital</th>\n",
       "      <th>DateOfConfirmation</th>\n",
       "      <th>Symptoms</th>\n",
       "      <th>LivesInWuhan</th>\n",
       "      <th>LivesInWuhanComment</th>\n",
       "      <th>TravelHistoryDates</th>\n",
       "      <th>ReportedMarketExposure</th>\n",
       "      <th>...</th>\n",
       "      <th>Entity[\"Country\", \"VaticanCity\"]</th>\n",
       "      <th>Entity[\"Country\", \"Venezuela\"]</th>\n",
       "      <th>Entity[\"Country\", \"Vietnam\"]</th>\n",
       "      <th>Entity[\"Country\", \"Zambia\"]</th>\n",
       "      <th>Entity[\"Country\", \"Zimbabwe\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Africa\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Asia\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Europe\"]</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>48</td>\n",
       "      <td>Entity[\"Gender\", \"Female\"]</td>\n",
       "      <td>DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\", \"fatigue\", \"fever\", \"sputum\", \"myalg...</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Interval[{15, 88}]</td>\n",
       "      <td>Entity[\"Gender\", \"Female\"]</td>\n",
       "      <td>DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"shortness of breath\", \"respiratory symptoms\"}</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>61</td>\n",
       "      <td>Entity[\"Gender\", \"Male\"]</td>\n",
       "      <td>DateObject[{2019, 12, 20}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 27}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>58</td>\n",
       "      <td>Entity[\"Gender\", \"Male\"]</td>\n",
       "      <td>DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2020, 1, 1}, \"Day\", \"Gregorian\", -5.]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\", \"sputum\", \"fever\"}</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>66</td>\n",
       "      <td>Entity[\"Gender\", \"Male\"]</td>\n",
       "      <td>DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\"}</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 382 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Age                         Sex  \\\n",
       "0                  48  Entity[\"Gender\", \"Female\"]   \n",
       "1  Interval[{15, 88}]  Entity[\"Gender\", \"Female\"]   \n",
       "2                  61    Entity[\"Gender\", \"Male\"]   \n",
       "3                  58    Entity[\"Gender\", \"Male\"]   \n",
       "4                  66    Entity[\"Gender\", \"Male\"]   \n",
       "\n",
       "                                 DateOfOnsetSymptoms  \\\n",
       "0  DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...   \n",
       "1  DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...   \n",
       "2  DateObject[{2019, 12, 20}, \"Day\", \"Gregorian\",...   \n",
       "3  DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...   \n",
       "4  DateObject[{2019, 12, 22}, \"Day\", \"Gregorian\",...   \n",
       "\n",
       "                             DateOfAdmissionHospital       DateOfConfirmation  \\\n",
       "0  DateObject[{2019, 12, 10}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "1  DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "2  DateObject[{2019, 12, 27}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "3  DateObject[{2020, 1, 1}, \"Day\", \"Gregorian\", -5.]  Missing[\"NotAvailable\"]   \n",
       "4  DateObject[{2019, 12, 31}, \"Day\", \"Gregorian\",...  Missing[\"NotAvailable\"]   \n",
       "\n",
       "                                            Symptoms             LivesInWuhan  \\\n",
       "0  {\"cough\", \"fatigue\", \"fever\", \"sputum\", \"myalg...                     True   \n",
       "1    {\"shortness of breath\", \"respiratory symptoms\"}  Missing[\"NotAvailable\"]   \n",
       "2                            Missing[\"NotAvailable\"]                     True   \n",
       "3                       {\"cough\", \"sputum\", \"fever\"}                     True   \n",
       "4                                          {\"cough\"}                     True   \n",
       "\n",
       "       LivesInWuhanComment       TravelHistoryDates   ReportedMarketExposure  \\\n",
       "0  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "1  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "3  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "4  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "\n",
       "   ... Entity[\"Country\", \"VaticanCity\"]  Entity[\"Country\", \"Venezuela\"]  \\\n",
       "0  ...                            False                           False   \n",
       "1  ...                            False                           False   \n",
       "2  ...                            False                           False   \n",
       "3  ...                            False                           False   \n",
       "4  ...                            False                           False   \n",
       "\n",
       "  Entity[\"Country\", \"Vietnam\"] Entity[\"Country\", \"Zambia\"]  \\\n",
       "0                        False                       False   \n",
       "1                        False                       False   \n",
       "2                        False                       False   \n",
       "3                        False                       False   \n",
       "4                        False                       False   \n",
       "\n",
       "  Entity[\"Country\", \"Zimbabwe\"] Entity[\"GeographicRegion\", \"Africa\"]  \\\n",
       "0                         False                                False   \n",
       "1                         False                                False   \n",
       "2                         False                                False   \n",
       "3                         False                                False   \n",
       "4                         False                                False   \n",
       "\n",
       "  Entity[\"GeographicRegion\", \"Asia\"] Entity[\"GeographicRegion\", \"Europe\"]  \\\n",
       "0                              False                                False   \n",
       "1                              False                                False   \n",
       "2                              False                                False   \n",
       "3                              False                                False   \n",
       "4                              False                                False   \n",
       "\n",
       "   Latitude  Longitude  \n",
       "0  30.62506   114.3421  \n",
       "1  30.62506   114.3421  \n",
       "2  30.62506   114.3421  \n",
       "3  30.62506   114.3421  \n",
       "4  30.62506   114.3421  \n",
       "\n",
       "[5 rows x 382 columns]"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loc_dropped = df.drop(columns=['City', 'AdministrativeDivision', 'Country', 'GeoPosition'], axis=1)\n",
    "df = pd.concat([loc_dropped, latlong_frame], axis=1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Transforming the age entities and dealing with age intervals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "def age_transform(data):\n",
    "    numbers = re.findall(r'[0-9]+', data)\n",
    "    numbers = [int(x) for x in numbers]\n",
    "    if len(numbers) == 1:\n",
    "        return numbers[0]\n",
    "    # age interval\n",
    "    return sum(numbers) >> 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f85fdac0c90>"
      ]
     },
     "execution_count": 95,
     "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": [
    "df['Age'] = df['Age'].apply(age_transform)\n",
    "df['Age'].hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Transforming gender entities and filling missing data such that it mantains the initial distribution(where missing data is not counted) because it is a known fact that the disease is more prevalent among men(56% cases in the initial distribution)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "def gender_transform(data):\n",
    "    match = re.search(r\"Male|Female\", data)\n",
    "    if not match:\n",
    "        return None\n",
    "    return match.group()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Sex'] = df['Sex'].apply(gender_transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Male      0.560704\n",
      "Female    0.439296\n",
      "Name: Sex, dtype: float64\n",
      "\n",
      "count     17330\n",
      "unique        2\n",
      "top        Male\n",
      "freq       9717\n",
      "Name: Sex, dtype: object\n"
     ]
    }
   ],
   "source": [
    "counts = df['Sex'].value_counts(normalize=True)\n",
    "print(counts, end=\"\\n\\n\")\n",
    "print(df['Sex'].describe())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of missing gender data: 2137\n"
     ]
    }
   ],
   "source": [
    "missing = df['Sex'].isnull()\n",
    "missing_size =  len(missing.index[missing == True])\n",
    "print('Number of missing gender data:', missing_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "# fill missing data and mantain the distribution\n",
    "fillers = ['Male' if np.random.random() <  counts.values[0] else 'Female' for _ in range(missing_size)]\n",
    "df.loc[missing, 'Sex'] = fillers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Male      0.561052\n",
      "Female    0.438948\n",
      "Name: Sex, dtype: float64\n",
      "\n",
      "count     19467\n",
      "unique        2\n",
      "top        Male\n",
      "freq      10922\n",
      "Name: Sex, dtype: object\n"
     ]
    }
   ],
   "source": [
    "print(df['Sex'].value_counts(normalize=True), end=\"\\n\\n\")\n",
    "print(df['Sex'].describe())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Missing[\"NotAvailable\"]', 'True', 'False'], dtype=object)"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['ReportedMarketExposure'].unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Computing delta(days) between DateOfAdmissionHospital and DateOfOnsetSymptoms\n",
    "because data is more representative than actual dates.\n",
    "\n",
    "Filling in missing values with the mean."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "def date_transform(entity):\n",
    "    pattern = r\"[0-9]+\"\n",
    "    found = [int(x) for x in re.findall(pattern, entity)]\n",
    "    if re.search(r'^DateInterval', entity):\n",
    "        try:\n",
    "            # try to find midpoint\n",
    "            start = datetime.date(*found[0:3])\n",
    "            end   = datetime.date(*found[3:6])\n",
    "            mid = start + (end - start) / 2\n",
    "            return mid\n",
    "        except:\n",
    "            # try and catch because of examples such this one...\n",
    "            # DateInterval[{{{-Infinity, 1, 1}, {2020, 2, 25}}}, \"Day\", \"Gregorian\", -5.]\n",
    "            # i mean wtf\n",
    "            return np.nan\n",
    "    \n",
    "    if len(found) == 0: # missing data\n",
    "        return np.nan\n",
    "    # unpacking array to function arguments\n",
    "    date = datetime.date(*found[:3])\n",
    "    return date"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "count    1111.000000\n",
      "mean        5.295230\n",
      "std         4.880847\n",
      "min         1.000000\n",
      "25%         2.000000\n",
      "50%         4.000000\n",
      "75%         7.000000\n",
      "max        50.000000\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "adms_col  = 'DateOfAdmissionHospital'\n",
    "onset_col = 'DateOfOnsetSymptoms'\n",
    "delta_col = 'SymptomsAdmissionDelta'\n",
    "\n",
    "delta_df = df[adms_col].apply(date_transform) - df[onset_col].apply(date_transform)\n",
    "delta_df = delta_df.apply(lambda x: abs(x.days) if x else np.nan)\n",
    "print(delta_df.describe())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f86074fcd90>"
      ]
     },
     "execution_count": 105,
     "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": [
    "delta_df.hist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "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>Sex</th>\n",
       "      <th>DateOfConfirmation</th>\n",
       "      <th>Symptoms</th>\n",
       "      <th>LivesInWuhan</th>\n",
       "      <th>LivesInWuhanComment</th>\n",
       "      <th>TravelHistoryDates</th>\n",
       "      <th>ReportedMarketExposure</th>\n",
       "      <th>ReportedMarketExposureComment</th>\n",
       "      <th>ChronicDiseaseQ</th>\n",
       "      <th>...</th>\n",
       "      <th>Entity[\"Country\", \"Venezuela\"]</th>\n",
       "      <th>Entity[\"Country\", \"Vietnam\"]</th>\n",
       "      <th>Entity[\"Country\", \"Zambia\"]</th>\n",
       "      <th>Entity[\"Country\", \"Zimbabwe\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Africa\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Asia\"]</th>\n",
       "      <th>Entity[\"GeographicRegion\", \"Europe\"]</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "      <th>SymptomsAdmissionDelta</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>48</td>\n",
       "      <td>Female</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\", \"fatigue\", \"fever\", \"sputum\", \"myalg...</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "      <td>5.29523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>51</td>\n",
       "      <td>Female</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"shortness of breath\", \"respiratory symptoms\"}</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "      <td>21.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>61</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "      <td>7.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>58</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\", \"sputum\", \"fever\"}</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "      <td>10.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>66</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>{\"cough\"}</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>30.62506</td>\n",
       "      <td>114.3421</td>\n",
       "      <td>9.00000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 381 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Age     Sex       DateOfConfirmation  \\\n",
       "0   48  Female  Missing[\"NotAvailable\"]   \n",
       "1   51  Female  Missing[\"NotAvailable\"]   \n",
       "2   61    Male  Missing[\"NotAvailable\"]   \n",
       "3   58    Male  Missing[\"NotAvailable\"]   \n",
       "4   66    Male  Missing[\"NotAvailable\"]   \n",
       "\n",
       "                                            Symptoms             LivesInWuhan  \\\n",
       "0  {\"cough\", \"fatigue\", \"fever\", \"sputum\", \"myalg...                     True   \n",
       "1    {\"shortness of breath\", \"respiratory symptoms\"}  Missing[\"NotAvailable\"]   \n",
       "2                            Missing[\"NotAvailable\"]                     True   \n",
       "3                       {\"cough\", \"sputum\", \"fever\"}                     True   \n",
       "4                                          {\"cough\"}                     True   \n",
       "\n",
       "       LivesInWuhanComment       TravelHistoryDates   ReportedMarketExposure  \\\n",
       "0  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "1  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "3  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "4  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "\n",
       "  ReportedMarketExposureComment  ChronicDiseaseQ  ...  \\\n",
       "0       Missing[\"NotAvailable\"]             True  ...   \n",
       "1       Missing[\"NotAvailable\"]            False  ...   \n",
       "2       Missing[\"NotAvailable\"]            False  ...   \n",
       "3       Missing[\"NotAvailable\"]            False  ...   \n",
       "4       Missing[\"NotAvailable\"]            False  ...   \n",
       "\n",
       "  Entity[\"Country\", \"Venezuela\"] Entity[\"Country\", \"Vietnam\"]  \\\n",
       "0                          False                        False   \n",
       "1                          False                        False   \n",
       "2                          False                        False   \n",
       "3                          False                        False   \n",
       "4                          False                        False   \n",
       "\n",
       "  Entity[\"Country\", \"Zambia\"] Entity[\"Country\", \"Zimbabwe\"]  \\\n",
       "0                       False                         False   \n",
       "1                       False                         False   \n",
       "2                       False                         False   \n",
       "3                       False                         False   \n",
       "4                       False                         False   \n",
       "\n",
       "  Entity[\"GeographicRegion\", \"Africa\"] Entity[\"GeographicRegion\", \"Asia\"]  \\\n",
       "0                                False                              False   \n",
       "1                                False                              False   \n",
       "2                                False                              False   \n",
       "3                                False                              False   \n",
       "4                                False                              False   \n",
       "\n",
       "   Entity[\"GeographicRegion\", \"Europe\"]  Latitude  Longitude  \\\n",
       "0                                 False  30.62506   114.3421   \n",
       "1                                 False  30.62506   114.3421   \n",
       "2                                 False  30.62506   114.3421   \n",
       "3                                 False  30.62506   114.3421   \n",
       "4                                 False  30.62506   114.3421   \n",
       "\n",
       "   SymptomsAdmissionDelta  \n",
       "0                 5.29523  \n",
       "1                21.00000  \n",
       "2                 7.00000  \n",
       "3                10.00000  \n",
       "4                 9.00000  \n",
       "\n",
       "[5 rows x 381 columns]"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "delta_df = delta_df.fillna(delta_df.mean())\n",
    "df[delta_col] = delta_df\n",
    "df = df.drop(columns=[adms_col, onset_col])\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One hot encoding for symptoms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "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>abdominal pain</th>\n",
       "      <th>acute coronary syndrome</th>\n",
       "      <th>acute kidney injury</th>\n",
       "      <th>acute left heart failure</th>\n",
       "      <th>acute myocardial infarction</th>\n",
       "      <th>acute pharyngitis</th>\n",
       "      <th>acute renal failure</th>\n",
       "      <th>acute respiratory disease</th>\n",
       "      <th>acute respiratory disease syndrome</th>\n",
       "      <th>acute respiratory distress syndrome</th>\n",
       "      <th>...</th>\n",
       "      <th>sore limbs</th>\n",
       "      <th>sore muscle</th>\n",
       "      <th>sore throat</th>\n",
       "      <th>sputum</th>\n",
       "      <th>sweating</th>\n",
       "      <th>systemic weakness</th>\n",
       "      <th>toothache</th>\n",
       "      <th>vomiting</th>\n",
       "      <th>weakness</th>\n",
       "      <th>wheezing</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 92 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   abdominal pain  acute coronary syndrome  acute kidney injury  \\\n",
       "0           False                    False                False   \n",
       "1           False                    False                False   \n",
       "2           False                    False                False   \n",
       "3           False                    False                False   \n",
       "4           False                    False                False   \n",
       "\n",
       "   acute left heart failure  acute myocardial infarction  acute pharyngitis  \\\n",
       "0                     False                        False              False   \n",
       "1                     False                        False              False   \n",
       "2                     False                        False              False   \n",
       "3                     False                        False              False   \n",
       "4                     False                        False              False   \n",
       "\n",
       "   acute renal failure  acute respiratory disease  \\\n",
       "0                False                      False   \n",
       "1                False                      False   \n",
       "2                False                      False   \n",
       "3                False                      False   \n",
       "4                False                      False   \n",
       "\n",
       "   acute respiratory disease syndrome  acute respiratory distress syndrome  \\\n",
       "0                               False                                False   \n",
       "1                               False                                False   \n",
       "2                               False                                False   \n",
       "3                               False                                False   \n",
       "4                               False                                False   \n",
       "\n",
       "   ...  sore limbs  sore muscle  sore throat  sputum  sweating  \\\n",
       "0  ...       False        False        False    True     False   \n",
       "1  ...       False        False        False   False     False   \n",
       "2  ...       False        False        False   False     False   \n",
       "3  ...       False        False        False    True     False   \n",
       "4  ...       False        False        False   False     False   \n",
       "\n",
       "   systemic weakness  toothache  vomiting  weakness  wheezing  \n",
       "0              False      False     False     False     False  \n",
       "1              False      False     False     False     False  \n",
       "2              False      False     False     False     False  \n",
       "3              False      False     False     False     False  \n",
       "4              False      False     False     False     False  \n",
       "\n",
       "[5 rows x 92 columns]"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# transform to list of symptoms\n",
    "transform = lambda x: re.findall(r\"\\\"([a-z ]+)\\\"\", x)\n",
    "symptoms_lists = df['Symptoms'].apply(transform)\n",
    "# get the unique symptoms found in the dataset\n",
    "symptoms = np.concatenate(symptoms_lists.values)\n",
    "symptoms = np.unique(symptoms)\n",
    "onehot = np.array([np.in1d(symptoms, list_col) for list_col in symptoms_lists])\n",
    "# create new dataframe using the unique symptoms as column headers\n",
    "onehot_frame = pd.DataFrame(onehot, columns=symptoms)\n",
    "onehot_frame.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f85fdabcbd0>"
      ]
     },
     "execution_count": 108,
     "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": [
    "sum_agg = onehot_frame.astype(float).agg('sum')\n",
    "sum_agg.hist(bins=92)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fever                                  1031.0\n",
       "cough                                   577.0\n",
       "mild symptoms                           184.0\n",
       "sore throat                             118.0\n",
       "pneumonia                               109.0\n",
       "fatigue                                  82.0\n",
       "malaise                                  79.0\n",
       "headache                                 74.0\n",
       "rhinorrhea                               73.0\n",
       "myalgias                                 69.0\n",
       "shortness of breath                      65.0\n",
       "sputum                                   49.0\n",
       "chills                                   41.0\n",
       "pneumonitis                              40.0\n",
       "acute respiratory distress syndrome      37.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum_agg.sort_values(ascending=False).head(15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "symptoms_dropped = df.drop(columns=['Symptoms'], axis=1)\n",
    "df = pd.concat([symptoms_dropped, onehot_frame], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "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>Sex</th>\n",
       "      <th>DateOfConfirmation</th>\n",
       "      <th>LivesInWuhan</th>\n",
       "      <th>LivesInWuhanComment</th>\n",
       "      <th>TravelHistoryDates</th>\n",
       "      <th>ReportedMarketExposure</th>\n",
       "      <th>ReportedMarketExposureComment</th>\n",
       "      <th>ChronicDiseaseQ</th>\n",
       "      <th>ChronicDiseases</th>\n",
       "      <th>...</th>\n",
       "      <th>sore limbs</th>\n",
       "      <th>sore muscle</th>\n",
       "      <th>sore throat</th>\n",
       "      <th>sputum</th>\n",
       "      <th>sweating</th>\n",
       "      <th>systemic weakness</th>\n",
       "      <th>toothache</th>\n",
       "      <th>vomiting</th>\n",
       "      <th>weakness</th>\n",
       "      <th>wheezing</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>48</td>\n",
       "      <td>Female</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>{\"diabetes\", \"cerebral infarction\"}</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>51</td>\n",
       "      <td>Female</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>61</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>58</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>66</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>False</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 472 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Age     Sex       DateOfConfirmation             LivesInWuhan  \\\n",
       "0   48  Female  Missing[\"NotAvailable\"]                     True   \n",
       "1   51  Female  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2   61    Male  Missing[\"NotAvailable\"]                     True   \n",
       "3   58    Male  Missing[\"NotAvailable\"]                     True   \n",
       "4   66    Male  Missing[\"NotAvailable\"]                     True   \n",
       "\n",
       "       LivesInWuhanComment       TravelHistoryDates   ReportedMarketExposure  \\\n",
       "0  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "1  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "3  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "4  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "\n",
       "  ReportedMarketExposureComment  ChronicDiseaseQ  \\\n",
       "0       Missing[\"NotAvailable\"]             True   \n",
       "1       Missing[\"NotAvailable\"]            False   \n",
       "2       Missing[\"NotAvailable\"]            False   \n",
       "3       Missing[\"NotAvailable\"]            False   \n",
       "4       Missing[\"NotAvailable\"]            False   \n",
       "\n",
       "                       ChronicDiseases  ... sore limbs sore muscle  \\\n",
       "0  {\"diabetes\", \"cerebral infarction\"}  ...      False       False   \n",
       "1              Missing[\"NotAvailable\"]  ...      False       False   \n",
       "2              Missing[\"NotAvailable\"]  ...      False       False   \n",
       "3              Missing[\"NotAvailable\"]  ...      False       False   \n",
       "4              Missing[\"NotAvailable\"]  ...      False       False   \n",
       "\n",
       "  sore throat sputum sweating  systemic weakness  toothache  vomiting  \\\n",
       "0       False   True    False              False      False     False   \n",
       "1       False  False    False              False      False     False   \n",
       "2       False  False    False              False      False     False   \n",
       "3       False   True    False              False      False     False   \n",
       "4       False  False    False              False      False     False   \n",
       "\n",
       "   weakness  wheezing  \n",
       "0     False     False  \n",
       "1     False     False  \n",
       "2     False     False  \n",
       "3     False     False  \n",
       "4     False     False  \n",
       "\n",
       "[5 rows x 472 columns]"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One hot encoding for chronic diseases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "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>asthma</th>\n",
       "      <th>atherosclerosis</th>\n",
       "      <th>atrial fibrillation</th>\n",
       "      <th>benign prostatic hyperplasia</th>\n",
       "      <th>benign prostatic hypertrophy</th>\n",
       "      <th>bronchial asthma</th>\n",
       "      <th>cardiac disease</th>\n",
       "      <th>cardiac dysrhythmia</th>\n",
       "      <th>cardiomyopathy</th>\n",
       "      <th>cardiovascular disease</th>\n",
       "      <th>...</th>\n",
       "      <th>on dialysis</th>\n",
       "      <th>prerenal azotemia</th>\n",
       "      <th>prostate cancer</th>\n",
       "      <th>prostate hypertrophy</th>\n",
       "      <th>stenocardia</th>\n",
       "      <th>systemic arterial hypertension</th>\n",
       "      <th>tongue cancer</th>\n",
       "      <th>tuberculosis</th>\n",
       "      <th>upper gastrointestinal bleeding</th>\n",
       "      <th>valvular heart disease</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 49 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   asthma  atherosclerosis  atrial fibrillation  benign prostatic hyperplasia  \\\n",
       "0   False            False                False                         False   \n",
       "1   False            False                False                         False   \n",
       "2   False            False                False                         False   \n",
       "3   False            False                False                         False   \n",
       "4   False            False                False                         False   \n",
       "\n",
       "   benign prostatic hypertrophy  bronchial asthma  cardiac disease  \\\n",
       "0                         False             False            False   \n",
       "1                         False             False            False   \n",
       "2                         False             False            False   \n",
       "3                         False             False            False   \n",
       "4                         False             False            False   \n",
       "\n",
       "   cardiac dysrhythmia  cardiomyopathy  cardiovascular disease  ...  \\\n",
       "0                False           False                   False  ...   \n",
       "1                False           False                   False  ...   \n",
       "2                False           False                   False  ...   \n",
       "3                False           False                   False  ...   \n",
       "4                False           False                   False  ...   \n",
       "\n",
       "   on dialysis  prerenal azotemia  prostate cancer  prostate hypertrophy  \\\n",
       "0        False              False            False                 False   \n",
       "1        False              False            False                 False   \n",
       "2        False              False            False                 False   \n",
       "3        False              False            False                 False   \n",
       "4        False              False            False                 False   \n",
       "\n",
       "   stenocardia  systemic arterial hypertension  tongue cancer  tuberculosis  \\\n",
       "0        False                           False          False         False   \n",
       "1        False                           False          False         False   \n",
       "2        False                           False          False         False   \n",
       "3        False                           False          False         False   \n",
       "4        False                           False          False         False   \n",
       "\n",
       "   upper gastrointestinal bleeding  valvular heart disease  \n",
       "0                            False                   False  \n",
       "1                            False                   False  \n",
       "2                            False                   False  \n",
       "3                            False                   False  \n",
       "4                            False                   False  \n",
       "\n",
       "[5 rows x 49 columns]"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# transform to list of symptoms\n",
    "transform = lambda x: re.findall(r\"\\\"([a-z ]+)\\\"\", x)\n",
    "diseases_lists = df['ChronicDiseases'].apply(transform)\n",
    "# get the unique symptoms found in the dataset\n",
    "diseases = np.concatenate(diseases_lists.values)\n",
    "diseases = np.unique(diseases)\n",
    "onehot = np.array([np.in1d(diseases, list_col) for list_col in diseases_lists])\n",
    "# create new dataframe using the unique diseases as column headers\n",
    "onehot_frame = pd.DataFrame(onehot, columns=diseases)\n",
    "onehot_frame.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8605e44c90>"
      ]
     },
     "execution_count": 113,
     "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": [
    "sum_agg = onehot_frame.astype(float).agg('sum')\n",
    "sum_agg.hist(bins=49)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "hypertension                             100.0\n",
       "diabetes                                  70.0\n",
       "chronic kidney disease                    13.0\n",
       "asthma                                     9.0\n",
       "coronary heart disease                     9.0\n",
       "chronic obstructive pulmonary disease      7.0\n",
       "cardiac disease                            5.0\n",
       "kidney disease                             4.0\n",
       "chronic bronchitis                         4.0\n",
       "coronary stent                             4.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum_agg.sort_values(ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [],
   "source": [
    "diseases_dropped = df.drop(columns=['ChronicDiseases', 'ChronicDiseaseQ'], axis=1)\n",
    "df = pd.concat([diseases_dropped, onehot_frame], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Age</th>\n",
       "      <th>Sex</th>\n",
       "      <th>DateOfConfirmation</th>\n",
       "      <th>LivesInWuhan</th>\n",
       "      <th>LivesInWuhanComment</th>\n",
       "      <th>TravelHistoryDates</th>\n",
       "      <th>ReportedMarketExposure</th>\n",
       "      <th>ReportedMarketExposureComment</th>\n",
       "      <th>SequenceAvailable</th>\n",
       "      <th>DischargedQ</th>\n",
       "      <th>...</th>\n",
       "      <th>on dialysis</th>\n",
       "      <th>prerenal azotemia</th>\n",
       "      <th>prostate cancer</th>\n",
       "      <th>prostate hypertrophy</th>\n",
       "      <th>stenocardia</th>\n",
       "      <th>systemic arterial hypertension</th>\n",
       "      <th>tongue cancer</th>\n",
       "      <th>tuberculosis</th>\n",
       "      <th>upper gastrointestinal bleeding</th>\n",
       "      <th>valvular heart disease</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>48</td>\n",
       "      <td>Female</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>51</td>\n",
       "      <td>Female</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>61</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>58</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>66</td>\n",
       "      <td>Male</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>True</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>Missing[\"NotAvailable\"]</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 519 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Age     Sex       DateOfConfirmation             LivesInWuhan  \\\n",
       "0   48  Female  Missing[\"NotAvailable\"]                     True   \n",
       "1   51  Female  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2   61    Male  Missing[\"NotAvailable\"]                     True   \n",
       "3   58    Male  Missing[\"NotAvailable\"]                     True   \n",
       "4   66    Male  Missing[\"NotAvailable\"]                     True   \n",
       "\n",
       "       LivesInWuhanComment       TravelHistoryDates   ReportedMarketExposure  \\\n",
       "0  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "1  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "3  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "4  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "\n",
       "  ReportedMarketExposureComment        SequenceAvailable  \\\n",
       "0       Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "1       Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "2       Missing[\"NotAvailable\"]                     True   \n",
       "3       Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "4       Missing[\"NotAvailable\"]  Missing[\"NotAvailable\"]   \n",
       "\n",
       "               DischargedQ  ... on dialysis prerenal azotemia prostate cancer  \\\n",
       "0  Missing[\"NotAvailable\"]  ...       False             False           False   \n",
       "1  Missing[\"NotAvailable\"]  ...       False             False           False   \n",
       "2  Missing[\"NotAvailable\"]  ...       False             False           False   \n",
       "3  Missing[\"NotAvailable\"]  ...       False             False           False   \n",
       "4  Missing[\"NotAvailable\"]  ...       False             False           False   \n",
       "\n",
       "   prostate hypertrophy  stenocardia  systemic arterial hypertension  \\\n",
       "0                 False        False                           False   \n",
       "1                 False        False                           False   \n",
       "2                 False        False                           False   \n",
       "3                 False        False                           False   \n",
       "4                 False        False                           False   \n",
       "\n",
       "   tongue cancer  tuberculosis  upper gastrointestinal bleeding  \\\n",
       "0          False         False                            False   \n",
       "1          False         False                            False   \n",
       "2          False         False                            False   \n",
       "3          False         False                            False   \n",
       "4          False         False                            False   \n",
       "\n",
       "   valvular heart disease  \n",
       "0                   False  \n",
       "1                   False  \n",
       "2                   False  \n",
       "3                   False  \n",
       "4                   False  \n",
       "\n",
       "[5 rows x 519 columns]"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop useless features\n",
    "df = df.drop(columns=['LivesInWuhanComment', 'TravelHistoryDates', \n",
    "                      'ReportedMarketExposureComment', 'SequenceAvailable',\n",
    "                      'DateOfConfirmation', 'DischargedQ',\n",
    "                      'DateOfDeath', 'DateOfDischarge'], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Replace missing binary data with false"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "replacer = {'Missing[\"NotAvailable\"]': False, 'False': False, 'True': True}\n",
    "columns = ['LivesInWuhan', 'ReportedMarketExposure', 'DeathQ']\n",
    "\n",
    "for column in columns:\n",
    "    df[column] = df[column].replace(replacer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Encode gender column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Sex'] = df['Sex'].replace({'Male': False, 'Female': True})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preprocessed dataset:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Age</th>\n",
       "      <th>Sex</th>\n",
       "      <th>LivesInWuhan</th>\n",
       "      <th>ReportedMarketExposure</th>\n",
       "      <th>DeathQ</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"AbuDhabi\", \"UnitedArabEmirates\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Accra\", \"GreaterAccra\", \"Ghana\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"AddisAbaba\", \"Ethiopia\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Aichi\", \"Japan\"}]</th>\n",
       "      <th>Entity[\"AdministrativeDivision\", {\"Alabama\", \"UnitedStates\"}]</th>\n",
       "      <th>...</th>\n",
       "      <th>on dialysis</th>\n",
       "      <th>prerenal azotemia</th>\n",
       "      <th>prostate cancer</th>\n",
       "      <th>prostate hypertrophy</th>\n",
       "      <th>stenocardia</th>\n",
       "      <th>systemic arterial hypertension</th>\n",
       "      <th>tongue cancer</th>\n",
       "      <th>tuberculosis</th>\n",
       "      <th>upper gastrointestinal bleeding</th>\n",
       "      <th>valvular heart disease</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>48</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</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",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>51</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</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",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>61</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</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",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>58</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</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",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>66</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</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 × 511 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Age  Sex  LivesInWuhan  ReportedMarketExposure  DeathQ  \\\n",
       "0   48    1             1                       0       1   \n",
       "1   51    1             0                       0       1   \n",
       "2   61    0             1                       0       1   \n",
       "3   58    0             1                       0       1   \n",
       "4   66    0             1                       0       1   \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"AbuDhabi\", \"UnitedArabEmirates\"}]  \\\n",
       "0                                                  0                      \n",
       "1                                                  0                      \n",
       "2                                                  0                      \n",
       "3                                                  0                      \n",
       "4                                                  0                      \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Accra\", \"GreaterAccra\", \"Ghana\"}]  \\\n",
       "0                                                  0                      \n",
       "1                                                  0                      \n",
       "2                                                  0                      \n",
       "3                                                  0                      \n",
       "4                                                  0                      \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"AddisAbaba\", \"Ethiopia\"}]  \\\n",
       "0                                                  0              \n",
       "1                                                  0              \n",
       "2                                                  0              \n",
       "3                                                  0              \n",
       "4                                                  0              \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Aichi\", \"Japan\"}]  \\\n",
       "0                                                  0      \n",
       "1                                                  0      \n",
       "2                                                  0      \n",
       "3                                                  0      \n",
       "4                                                  0      \n",
       "\n",
       "   Entity[\"AdministrativeDivision\", {\"Alabama\", \"UnitedStates\"}]  ...  \\\n",
       "0                                                  0              ...   \n",
       "1                                                  0              ...   \n",
       "2                                                  0              ...   \n",
       "3                                                  0              ...   \n",
       "4                                                  0              ...   \n",
       "\n",
       "   on dialysis  prerenal azotemia  prostate cancer  prostate hypertrophy  \\\n",
       "0            0                  0                0                     0   \n",
       "1            0                  0                0                     0   \n",
       "2            0                  0                0                     0   \n",
       "3            0                  0                0                     0   \n",
       "4            0                  0                0                     0   \n",
       "\n",
       "   stenocardia  systemic arterial hypertension  tongue cancer  tuberculosis  \\\n",
       "0            0                               0              0             0   \n",
       "1            0                               0              0             0   \n",
       "2            0                               0              0             0   \n",
       "3            0                               0              0             0   \n",
       "4            0                               0              0             0   \n",
       "\n",
       "   upper gastrointestinal bleeding  valvular heart disease  \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 511 columns]"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Preprocessed dataset:')\n",
    "# Convert from boolean to int notation\n",
    "df = df * 1\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "correlation_mat = df.corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As expected, the features are not strongly correlated, apart from chronic diseases/symptoms as seen in the bottom right corner(they were the last ones appended to the dataframe)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f8605b39390>"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(correlation_mat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_x = df.drop('DeathQ', axis=1).copy()\n",
    "df_y = df['DeathQ'].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "minmax_scaling = lambda x: (x - x.min()) / (x.max() - x.min())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Scale number features between 0 and 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [],
   "source": [
    "nbr_cols = ['Age', 'SymptomsAdmissionDelta', 'Latitude', 'Longitude']\n",
    "\n",
    "for col in nbr_cols:\n",
    "    df_x[col] = minmax_scaling(df_x[col])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import cross_val_score, KFold, train_test_split, validation_curve\n",
    "from sklearn import metrics\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   1 out of   1 | elapsed:   10.1s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=5.5, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='auto', n_jobs=-1, penalty='l2', random_state=42,\n",
       "                   solver='lbfgs', tol=0.0001, verbose=1, warm_start=False)"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df_x.values\n",
    "y = df_y.values\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "model = LogisticRegression(n_jobs=-1, verbose=1, random_state=42, max_iter=100,  penalty='l2', C=5.5)\n",
    "model.fit(X_train, y_train)\n",
    "\n",
    "#kfolding = KFold(n_splits=10, shuffle=True, random_state=42)\n",
    "#score = cross_val_score(model, X, y, n_jobs=4, cv=kfolding, scoring=\"accuracy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9961479198767335"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict_proba(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.018784914509306198"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.log_loss(y_test, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8605a4d890>"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWcAAAD4CAYAAAAw/yevAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAWnElEQVR4nO3df5RU5Z3n8fcHEMVfAUUJ0BhMhhlHdyM6KBrjj2hEUHcwOesezC+OYU87CW5+jJsVk534azxrMjruOkPcNBEFTSScREeO42hYkmySzYi0igREQy+KNKCMgibqBuiq7/5RD6YGuquroZp6+ubz4jynqr73qXufe06fL9/z3KfuVURgZmZ5GdTsAZiZ2d6cnM3MMuTkbGaWISdnM7MMOTmbmWVoSH8fYNdr670cxPYybMzZzR6CZahr5ybt7z76knMOGvn+/T5ef3HlbGaWoX6vnM3MDqhyqdkjaAgnZzMrllJXs0fQEE7OZlYoEeVmD6EhnJzNrFjKTs5mZvlx5WxmliFfEDQzy5ArZzOz/IRXa5iZZcgXBM3MMuRpDTOzDPmCoJlZhlw5m5llyBcEzcwy5AuCZmb5ifCcs5lZfgoy5+yb7ZtZsZTL9bcaJB0i6UlJz0paI+nGFL9X0ouSVqY2McUl6U5JHZJWSTq1al8zJa1LbWY9p+HK2cyKpXGV8w7g/Ih4S9JBwC8k/VPa9pWI+MEe/acBE1KbDNwFTJZ0FHA9MAkI4ClJSyJie62DOzmbWbGUdjVkNxERwFvp40Gp1Xo+4XRgYfreE5KGSxoNnAcsjYhtAJKWAlOBB2od39MaZlYsfZjWkNQqqb2qtVbvStJgSSuBrVQS7PK06ZY0dXGHpINTbCywserrnSnWU7wmV85mVix9mNaIiDagrcb2EjBR0nDgIUn/BrgOeAUYmr57LXAT0N2TvKNGvCZXzmZWLA26IFgtIt4AfgpMjYgtUbEDuAc4PXXrBMZVfa0F2FwjXpOTs5kVS+NWaxyTKmYkDQM+Cjyf5pGRJOAyYHX6yhLgM2nVxhnAmxGxBXgcmCJphKQRwJQUq8nTGmZWKNGgC4LAaGCBpMFUCtnFEfGIpB9LOobKdMVK4C9S/0eBi4EO4B3gSoCI2CbpZmBF6nfT7ouDtTg5m1mxNGgpXUSsAk7pJn5+D/0DmN3DtvnA/L4c38nZzIrF99YwM8tQQX6+7eRsZsXiytnMLEOunM3MMtTlm+2bmeXHlbOZWYY852xmliFXzmZmGXLlbGaWIVfOZmYZ8moNM7MMRa+3Sh4QnJzNrFg852xmliEnZzOzDPmCoJlZhkqlZo+gIZyczaxYPK1hZpYhJ2czswwVZM7ZT982s0KJctTdapF0iKQnJT0raY2kG1P8eEnLJa2T9H1JQ1P84PS5I20fX7Wv61L8BUkX1XMeTs5mVizlcv2tth3A+RFxMjARmCrpDOAbwB0RMQHYDsxK/WcB2yPij4A7Uj8knQjMAE4CpgLfSk/0rsnJ2cyKpVSqv9UQFW+ljwelFsD5wA9SfAFwWXo/PX0mbb9AklJ8UUTsiIgXgQ7g9N5Ow8nZzIqlD5WzpFZJ7VWttXpXkgZLWglsBZYC/xd4IyJ238CjExib3o8FNgKk7W8CR1fHu/lOj3xB0MyKpQ+rNSKiDWirsb0ETJQ0HHgI+NPuuqVX9bCtp3hNTs77YceOncyc/RV27tpFqavEhR/5MFf/x0/zRPsz3D73bsrl4NBDD+GWr13DcS1jAHhs2c/41vz7EeJPJryfb95w7bv7e+vtt/nzT1zFBed8iK9d8/lmnZYdIPPabueSiz/K1n95jYmnXNDs4RRHP9z4KCLekPRT4AxguKQhqTpuATanbp3AOKBT0hDgPcC2qvhu1d/pkac19sPQoQcx/85beXDBt/jBgrn8n+VP8ezqtdx821xuvf6/8MMFc7nkwo/w7XsfAGDDxk18577vc99dt/Pwd7/NtV+86l/t7+/m3cekU/5tM07FmmDhwsVccuknmz2M4mnQBUFJx6SKGUnDgI8Ca4GfAP8+dZsJPJzeL0mfSdt/HBGR4jPSao7jgQnAk72dRq+Vs6QTqExoj6VSim8GlkTE2t6+W3SSOPTQYQB0dXXR1dWFJAS8/fY7APz2rbc5ZuTRAPxgyWPM+Pi/4z1HHgHA0SOGv7uvNc+v4/Vt2/nw5EmseX7dgT0Ra4qf/2I573tfS7OHUTy9LJHrg9HAgrSyYhCwOCIekfQcsEjSXwPPAHen/ncD90nqoFIxzwCIiDWSFgPPAV3A7DRdUlPN5CzpWuAKYBG/z/QtwAOSFkXErX071+IplUr8h89+gZc3beaKj1/KB086gRvnfInP/eevc8jBQznssEP5XtsdQKVyBvjUX1xDuVTi87M+xYfPmES5XOZv/n4e/+2vvsLy9pXNPB2zga9B99aIiFXAKd3E19PNaouI+B1weQ/7ugW4pS/H721aYxZwWkTcGhH3p3ZrGtisnr5UfQX0Owsf6Mt4BpzBgwfzwwVzWfbQffzquV+zbv1LLPz+Q9x1200s+4f7ueziKXzzznkAdJVKbOjcxD1//w2+eeMcrr/1v/Ob377Fogcf4ZwzT2P0qGOafDZmA1+Uy3W3nPU2rVEGxgAb9oiPTtu6VX0FdNdr64vxWIJeHHnE4Zx26gf5+T+380LHej540gkATLvgHK665r8CMOqYkZx80gkcNGQILWPey/jjWtjQuYlnV6/lqVVrWPTgI7zz/37Hrl27OPTQQ/jy5z7bzFMyG5gaN63RVL0l5y8ByySt4/fr9I4D/gi4uj8HNhBs2/4GQ4YM4cgjDud3O3bwxIpn+OynLuett9/hpZc7GX9cC79c8Qzvf99xAFxwzpk8uvSnXHbJhWx/401e2riJcWNG842qFRv/8I9LWfP8Oidms31VkHtr1EzOEfGYpD+mMo0xlsp6vU5gRT0T2kX3L69v52t/fRulcpkoBxedfzbnnTWZG679Al/+2i1okDjyiMO5+bovA3DW5D/jl08+zZ9/spXBgwZzzexZDH/PkU0+C2uW+++by7nnnMnIkUfx0vp2brzpNu65d1GzhzXwFaRyVvTzwxD/UKY1rG+GjTm72UOwDHXt3NTdDzb65O2vz6g75xx206L9Pl5/8Y9QzKxY/hCmNczMBpyCTGs4OZtZoeS+RK5eTs5mViyunM3MMuTkbGaWoQb9fLvZnJzNrFB6ezbgQOHkbGbF4uRsZpYhr9YwM8uQK2czsww5OZuZ5SdKntYwM8uPK2czs/wUZSmdn75tZsVSjvpbDZLGSfqJpLWS1kj6YorfIGmTpJWpXVz1neskdUh6QdJFVfGpKdYhaU49p+HK2cyKpXFTzl3ANRHxtKQjgKckLU3b7oiI26o7SzqRyhO3T6LyeL//lR5WAjAXuJD0sBJJSyLiuVoHd3I2s0KJrsZk54jYAmxJ738raS2VJ0L1ZDqwKCJ2AC9K6uD3T+nuSE/tRtKi1Ldmcva0hpkVS7n+JqlVUntVa+1ul5LGA6cAy1PoakmrJM2XNCLFxvL7Z61CpUoeWyNek5OzmRVKlKP+FtEWEZOqWtue+5N0OPBD4EsR8RvgLuADwEQqlfXtu7t2N5wa8Zo8rWFmxdLAZc6SDqKSmL8bEQ8CRMSrVdvnAY+kj53AuKqvtwCb0/ue4j1y5WxmhdKXyrkWSQLuBtZGxN9WxUdXdfsYsDq9XwLMkHSwpOOBCcCTwApggqTjJQ2lctFwSW/n4crZzIqlcZXzWcCngV9JWpliXwWukDSRytTES8BVABGxRtJiKhf6uoDZEVECkHQ18DgwGJgfEWt6O7gi+nfB9q7X1hdjRbg11LAxZzd7CJahrp2bupuf7ZPXLzm37pxz9D/+7/0+Xn9x5WxmhRLFuLWGk7OZFYyTs5lZflw5m5llyMnZzCxDUcr2Gl+fODmbWaG4cjYzy1CUXTmbmWXHlbOZWYYiXDmbmWXHlbOZWYbKXq1hZpYfXxA0M8uQk7OZWYb6+UabB4yTs5kViitnM7MMeSmdmVmGSl6tYWaWH1fOZmYZKsqcs5++bWaFElF/q0XSOEk/kbRW0hpJX0zxoyQtlbQuvY5IcUm6U1KHpFWSTq3a18zUf52kmfWch5OzmRVKlFV360UXcE1E/ClwBjBb0onAHGBZREwAlqXPANOACam1AndBJZkD1wOTgdOB63cn9FqcnM2sUErlQXW3WiJiS0Q8nd7/FlgLjAWmAwtStwXAZen9dGBhVDwBDJc0GrgIWBoR2yJiO7AUmNrbeTg5m1mh9GVaQ1KrpPaq1trdPiWNB04BlgOjImJL5VixBTg2dRsLbKz6WmeK9RSvyRcEzaxQyn1YrRERbUBbrT6SDgd+CHwpIn4j9bj/7jZEjXhNrpzNrFAiVHfrjaSDqCTm70bEgyn8apquIL1uTfFOYFzV11uAzTXiNTk5m1mhNHC1hoC7gbUR8bdVm5YAu1dczAQerop/Jq3aOAN4M017PA5MkTQiXQickmI19fu0xrAxZ/f3IczM3tWXaY1enAV8GviVpJUp9lXgVmCxpFnAy8DladujwMVAB/AOcCVARGyTdDOwIvW7KSK29XZwRT/fwmnI0LEFuUeUmfW3rp2b9juzLh/z8bpzzuTND2b7ixVfEDSzQilKNejkbGaF0sBpjaZycjazQvGNj8zMMlSQh287OZtZsUS3v/kYeJyczaxQujytYWaWH1fOZmYZ8pyzmVmGXDmbmWXIlbOZWYZKrpzNzPJTkOe7OjmbWbGUXTmbmeXHNz4yM8uQLwiamWWo3PMz/gYUJ2czK5RSswfQIE7OZlYoXq1hZpahoqzW8NO3zaxQog+tN5LmS9oqaXVV7AZJmyStTO3iqm3XSeqQ9IKki6riU1OsQ9Kces7DydnMCqWs+lsd7gWmdhO/IyImpvYogKQTgRnASek735I0WNJgYC4wDTgRuCL1rcnTGmZWKI1cShcRP5M0vs7u04FFEbEDeFFSB3B62tYREesBJC1KfZ+rtTNXzmZWKCXV3yS1Smqvaq11HuZqSavStMeIFBsLbKzq05liPcVrcnI2s0Ip96FFRFtETKpqbXUc4i7gA8BEYAtwe4p3N1ESNeI1eVrDzAqlv38hGBGv7n4vaR7wSPrYCYyr6toCbE7ve4r3yJWzmRVKqP62LySNrvr4MWD3So4lwAxJB0s6HpgAPAmsACZIOl7SUCoXDZf0dhxXzmZWKI2snCU9AJwHjJTUCVwPnCdpIpWpiZeAqwAiYo2kxVQu9HUBsyOilPZzNfA4MBiYHxFrej12RP/ew2nI0LFFuUmUmfWzrp2b9vsXJH837lN155z/tPH+bH+x4srZzArFP982M8uQbxlqZpYhJ2czswwV5SKXk7OZFYrnnM3MMuSb7ZuZZahckIkNJ2czKxRfEDQzy1Ax6mYnZzMrGFfOZmYZ6lIxamcnZzMrlGKkZidnMysYT2uYmWXIS+nMzDJUjNTs5GxmBeNpDTOzDJUKUjs7OZtZobhyNjPLUBSkcvbTt82sUMp9aL2RNF/SVkmrq2JHSVoqaV16HZHiknSnpA5JqySdWvWdman/Okkz6zkPJ+d+Mq/tdjZ3PsvKZ5btte0vv3wVXTs3cfTRI5owMmum7v4uvv5Xf8mGF9tpX/Ej2lf8iGlTz2/iCAe+MlF3q8O9wNQ9YnOAZRExAViWPgNMAyak1grcBZVkTuWp3ZOB04Hrdyf0Wpyc+8nChYu55NJP7hVvaRnDRy84hw0bOpswKmu2nv4u/sed85h02hQmnTaFf3rsx00YWXFEH1qv+4r4GbBtj/B0YEF6vwC4rCq+MCqeAIZLGg1cBCyNiG0RsR1Yyt4Jfy9Ozv3k579Yzrbtb+wVv/22G5jz1VuIKMa8mPVNT38X1jhdRN1tH42KiC0A6fXYFB8LbKzq15liPcVrcnI+gC699EI2bdrCqlXPNXsolpnPf+5Knn5qKfPabmf48Pc0ezgDWvThn6RWSe1VrXU/Dt3dA7KiRrymfU7Okq6sse3dEy6X397XQxTKsGGH8NU5X+CGG29r9lAsM//z2wv54xM+xJ9NmsIrr2zlb7759WYPaUDrywXBiGiLiElVra2OQ7yapitIr1tTvBMYV9WvBdhcI17T/lTON/a0ofqEBw06bD8OURwf+MB4xo8/jqfbl9Lx6ydoaRnNiuWPM2rUMc0emjXZ1q2vUS6XiQi+c/d3Oe20ic0e0oDWl8p5Hy0Bdq+4mAk8XBX/TFq1cQbwZpr2eByYImlEuhA4JcVqqrnOWdKqnjYBo3o/B9tt9ernGdNy8rufO379BJPPnMbrr29v4qgsB+9977G88kql+Lps+jTWrHmhySMa2Br5IxRJDwDnASMldVJZdXErsFjSLOBl4PLU/VHgYqADeAe4EiAitkm6GViR+t0UEXteZNxLbz9CGUXlSuOeGUTAL3vb+R+y+++by7nnnMnIkUfx0vp2brzpNu65d1Gzh2VN1t3fxbnnfoiTTz6RiGDDhk4+9/lrmz3MAa3UwIvtEXFFD5su6KZvALN72M98YH5fjq1aqwYk3Q3cExG/6Gbb9yLiE70dYMjQsV6WYGZ16dq5qbuLZ33yifd9rO6c870ND+338fpLzco5ImbV2NZrYjYzO9CK8vNt31vDzArFNz4yM8uQn4RiZpYhT2uYmWWokas1msnJ2cwKxdMaZmYZ8gVBM7MMec7ZzCxDntYwM8tQUe6V7uRsZoVScuVsZpYfT2uYmWXI0xpmZhly5WxmliEvpTMzy5B/vm1mliFPa5iZZcjJ2cwsQ0VZrTGo2QMwM2ukMlF3642klyT9StJKSe0pdpSkpZLWpdcRKS5Jd0rqkLRK0qn7cx5OzmZWKNGHf3X6SERMjIhJ6fMcYFlETACWpc8A04AJqbUCd+3PeTg5m1mhlKJcd9tH04EF6f0C4LKq+MKoeAIYLmn0vh7EydnMCiUi6m6SWiW1V7XWPXcH/EjSU1XbRkXElnSsLcCxKT4W2Fj13c4U2ye+IGhmhdKX1RoR0Qa01ehyVkRslnQssFTS8zX6qrtD1D2YPbhyNrNCaeScc0RsTq9bgYeA04FXd09XpNetqXsnMK7q6y3A5n09DydnMyuUckTdrRZJh0k6Yvd7YAqwGlgCzEzdZgIPp/dLgM+kVRtnAG/unv7YF57WMLNCaeC9NUYBD0mCSq78XkQ8JmkFsFjSLOBl4PLU/1HgYqADeAe4cn8O7uRsZoWyH6sw/pWIWA+c3E38deCCbuIBzG7IwXFyNrOC6W26YqBwcjazQvEtQ83MMuTK2cwsQ66czcwyVIpSs4fQEE7OZlYoRbllqJOzmRWKb7ZvZpYhV85mZhnyag0zswx5tYaZWYYa9fPtZnNyNrNC8ZyzmVmGPOdsZpYhV85mZhnyOmczswy5cjYzy5BXa5iZZcgXBM3MMuRpDTOzDPkXgmZmGXLlbGaWoaLMOaso/8sMBJJaI6Kt2eOwvPjvwrozqNkD+APT2uwBWJb8d2F7cXI2M8uQk7OZWYacnA8szytad/x3YXvxBUEzswy5cjYzy5CTs5lZhpycDxBJUyW9IKlD0pxmj8eaT9J8SVslrW72WCw/Ts4HgKTBwFxgGnAicIWkE5s7KsvAvcDUZg/C8uTkfGCcDnRExPqI2AksAqY3eUzWZBHxM2Bbs8dheXJyPjDGAhurPnemmJlZt5ycDwx1E/MaRjPrkZPzgdEJjKv63AJsbtJYzGwAcHI+MFYAEyQdL2koMANY0uQxmVnGnJwPgIjoAq4GHgfWAosjYk1zR2XNJukB4J+BP5HUKWlWs8dk+fDPt83MMuTK2cwsQ07OZmYZcnI2M8uQk7OZWYacnM3MMuTkbGaWISdnM7MM/X/oJnPFcz5d5wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model.predict(X_test)\n",
    "conf = metrics.confusion_matrix(y_test, y_pred)\n",
    "sns.heatmap(conf, annot=True, fmt='d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_f1_scores_4conf_mat(conf_mat):\n",
    "  table = []\n",
    "  n = conf_mat.shape[0]\n",
    "\n",
    "  for i in range(n):\n",
    "    correct = conf_mat[i, i]\n",
    "    precision = correct / conf_mat[:, i].sum()\n",
    "    recall    = correct / conf_mat[i, :].sum()\n",
    "    f1 = 2 * precision * recall / (precision + recall)\n",
    "    table.append([precision, recall, f1])\n",
    "  return np.array(table)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.99638989, 0.99974127, 0.99806277],\n",
       "       [0.9375    , 0.51724138, 0.66666667]])"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_f1_scores_4conf_mat(conf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Trying to unsample negative entries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.utils import resample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = pd.concat([pd.DataFrame(X_train), \n",
    "                      pd.DataFrame(y_train, columns=['y'])], \n",
    "                      axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [],
   "source": [
    "neg = df_train[df_train['y'] == 0]\n",
    "pos = df_train[df_train['y'] == 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [],
   "source": [
    "neg_downsampled = resample(neg,\n",
    " replace=True, # sample with replacement\n",
    " n_samples=len(pos) * 3,\n",
    " random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    363\n",
       "1    121\n",
       "Name: y, dtype: int64"
      ]
     },
     "execution_count": 138,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "downsampled = pd.concat([neg_downsampled, pos])\n",
    "downsampled['y'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   1 out of   1 | elapsed:    0.5s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=5.5, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='auto', n_jobs=-1, penalty='l2', random_state=42,\n",
       "                   solver='lbfgs', tol=0.0001, verbose=1, warm_start=False)"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train = downsampled.drop(columns=['y']).copy()\n",
    "y_train = downsampled['y'].copy()\n",
    "model2 = LogisticRegression(n_jobs=-1, verbose=1, random_state=42, max_iter=100,  penalty='l2', C=5.5)\n",
    "model2.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9912686183872624"
      ]
     },
     "execution_count": 140,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f85fd7011d0>"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model2.predict(X_test)\n",
    "conf = metrics.confusion_matrix(y_test, y_pred)\n",
    "sns.heatmap(conf, annot=True, fmt='d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.99817945, 0.99301423, 0.99559014],\n",
       "       [0.44897959, 0.75862069, 0.56410256]])"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_f1_scores_4conf_mat(conf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Trying different class weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   1 out of   1 | elapsed:    4.1s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.5, class_weight={0: 0.01, 1: 0.99}, dual=False,\n",
       "                   fit_intercept=True, intercept_scaling=1, l1_ratio=None,\n",
       "                   max_iter=100, multi_class='auto', n_jobs=-1, penalty='l2',\n",
       "                   random_state=42, solver='lbfgs', tol=0.0001, verbose=1,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df_x.values\n",
    "y = df_y.values\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, \n",
    "                                                    random_state=42)\n",
    "class_weights = {0: 1e-2, 1: 99e-2}\n",
    "model3 = LogisticRegression(n_jobs=-1, verbose=1, \n",
    "                           random_state=42, max_iter=100,  \n",
    "                           penalty='l2', C=1.5, class_weight=class_weights)\n",
    "model3.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9861325115562404"
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model3.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f85fd6a4c90>"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model3.predict(X_test)\n",
    "conf = metrics.confusion_matrix(y_test, y_pred)\n",
    "sns.heatmap(conf, annot=True, fmt='d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.99843055, 0.98758085, 0.99297607],\n",
       "       [0.32394366, 0.79310345, 0.46      ]])"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_f1_scores_4conf_mat(conf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Ensambling resampled datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = pd.concat([pd.DataFrame(X_train), \n",
    "                      pd.DataFrame(y_train, columns=['y'])], \n",
    "                      axis=1)\n",
    "                      \n",
    "neg = df_train[df_train['y'] == 0]\n",
    "pos = df_train[df_train['y'] == 1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Using the a 1:4 pos/neg ratio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Will ensemble 32.0 models\n"
     ]
    }
   ],
   "source": [
    "ratio = 4\n",
    "count = np.ceil(len(neg) / len(pos) / ratio)\n",
    "print(f'Will ensemble {count} models')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [],
   "source": [
    "neg_shuffled = neg.values\n",
    "np.random.shuffle(neg_shuffled)\n",
    "neg_partitions = np.array_split(neg_shuffled, count, axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32 (483, 511)\n"
     ]
    }
   ],
   "source": [
    "print(len(neg_partitions), neg_partitions[0].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [],
   "source": [
    "def best_partition_model(neg_partitions, pos):\n",
    "    best = (None, 1e-10)\n",
    "    for index, partition in enumerate(neg_partitions, start=0):\n",
    "        dataset = np.concatenate((partition, pos.values), axis=0)\n",
    "        # drop the y column\n",
    "        y = dataset[:,  -1]\n",
    "        X = dataset[:, :-1]\n",
    "        np.random.shuffle(X_train)\n",
    "        model = LogisticRegression(n_jobs=-1, verbose=0, random_state=42, \n",
    "                                   max_iter=100,  penalty='l2', C=2.5)\n",
    "        model.fit(X, y)\n",
    "\n",
    "        y_pred = model.predict(X_test)\n",
    "        conf = metrics.confusion_matrix(y_test, y_pred)\n",
    "        scores = compute_f1_scores_4conf_mat(conf)\n",
    "        recall = scores[1, 1] # recall score for positive class\n",
    "\n",
    "        if recall > best[1]:\n",
    "            best = (model, recall)\n",
    "        if index % 10 == 0:\n",
    "            print(f'Progress: {index} / {count}')\n",
    "    return best"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress: 0 / 32.0\n",
      "Progress: 10 / 32.0\n",
      "Progress: 20 / 32.0\n",
      "Progress: 30 / 32.0\n",
      "(LogisticRegression(C=2.5, class_weight=None, dual=False, fit_intercept=True,\n",
      "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
      "                   multi_class='auto', n_jobs=-1, penalty='l2', random_state=42,\n",
      "                   solver='lbfgs', tol=0.0001, verbose=0, warm_start=False), 0.7586206896551724)\n"
     ]
    }
   ],
   "source": [
    "best = best_partition_model(neg_partitions, pos)\n",
    "print(best)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9922958397534669"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best[0].score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f85fd6f5a50>"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "conf = metrics.confusion_matrix(y_test, best[0].predict(X_test))\n",
    "sns.heatmap(conf, annot=True, fmt='d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.99818135, 0.99404916, 0.99611097],\n",
       "       [0.48888889, 0.75862069, 0.59459459]])"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_f1_scores_4conf_mat(conf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Use K-Nearest-Neighbors for ensembling datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [],
   "source": [
    "kmeans = KMeans(n_clusters=31, random_state=42, \n",
    "                n_jobs=-1, max_iter=100, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KMeans(algorithm='auto', copy_x=True, init='k-means++', max_iter=100,\n",
       "       n_clusters=31, n_init=10, n_jobs=-1, precompute_distances='auto',\n",
       "       random_state=42, tol=0.0001, verbose=1)"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kmeans.fit(neg.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([ 1, 10,  5, ...,  3, 11,  3], dtype=int32), (15452,), 31)"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kmeans.labels_, kmeans.labels_.shape, kmeans.n_clusters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "partitions = [neg.values[np.in1d(kmeans.labels_, cluster), :] for cluster in range(kmeans.n_clusters)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(141, 511) (2050, 511) (1182, 511) (2671, 511) (562, 511) (149, 511) (1494, 511) (113, 511) (181, 511) (180, 511) (1775, 511) (1047, 511) (261, 511) (197, 511) (638, 511) (204, 511) (143, 511) (241, 511) (204, 511) (202, 511) (139, 511) (162, 511) (114, 511) (109, 511) (164, 511) (102, 511) (157, 511) (63, 511) (8, 511) (607, 511) (192, 511)\n"
     ]
    }
   ],
   "source": [
    "print(' '.join([str(x.shape) for x in partitions]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Progress: 0 / 32.0\n",
      "Progress: 10 / 32.0\n",
      "Progress: 20 / 32.0\n",
      "Progress: 30 / 32.0\n",
      "(LogisticRegression(C=2.5, class_weight=None, dual=False, fit_intercept=True,\n",
      "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
      "                   multi_class='auto', n_jobs=-1, penalty='l2', random_state=42,\n",
      "                   solver='lbfgs', tol=0.0001, verbose=0, warm_start=False), 0.7586206896551724)\n"
     ]
    }
   ],
   "source": [
    "best2 = best_partition_model(partitions, pos)\n",
    "print(best)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.036466358500256806"
      ]
     },
     "execution_count": 163,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best2[0].score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Perfect recall for positive results, horrible accuracy and f1 scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f85f0e16690>"
      ]
     },
     "execution_count": 164,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "conf = metrics.confusion_matrix(y_test, best2[0].predict(X_test))\n",
    "sns.heatmap(conf, annot=True, fmt='d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.        , 0.02923674, 0.05681247],\n",
       "       [0.00766993, 1.        , 0.0152231 ]])"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_f1_scores_4conf_mat(conf)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.7.6 64-bit ('si': conda)",
   "language": "python",
   "name": "python37664bitsicondab72af9aab1a745ac8e94d9ea4d382c78"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}