COVID-19 Risk Modeling in East Java Using Geographically Weighted Logistic Regression Approach

Efforts to resolve the COVID-19 pandemic are a common goal set out in the Sustainable Development Goals (SDGs). However, the rate of spread of COVID-19 in several countries is still high. To control the spread ofthe COVID-19 pandemic, various countries have implemented several strategic measures. In Indonesia, the government through the COVID-19 Handling Task Force has classified areas based on the level of risk ofthe spread of COVID-19. East Java as one of the major provinces is a province dominated by regencies and cities with a moderate level of risk of spreading COVID-19 (orange zone) and low (yellow zone) in Indonesia. Controlling the rate of spread ofthe COVID-19 pandemic is relatively more controlled in the yellow zone area. For this reason, research was carried out to create statistical modeling related to the level of risk of COVID-19 in each regency and city in East Java which was influenced by various indicators. In this study, we used factors affecting level of risk of COVID-19 namely the number of confirmed positive cases of COVID-19, the recovery rate, and the fatality rate. The method used in the analysis is spatial approach based on the Geographically Weighted Logistic Regression (GWLR) method. The results show that the percentages of regency and cities in East Java that the recovery rate and the number of COVID-19 positive confirmations factors significantly affect on the level of risk of COVID-19 are 78.95% and 18.42 %, respectively. The GWLR model has better classification accuracy of 84.21% than that of global logistic regression model of 76.30%. The opportunity for an area to have a low level of risk of spreading COVID-19 will increase as the recovery rate increases. Therefore, the government can encourage each region to increase the recovery rate to reduce the spread of COVID-19.