Prediction of COVID-19 Spreading Using Support Vector Regression and Susceptible Infectious Recovered Model

Abstract

Many COVID-19 spread predictions have been implemented using various method. However, most of the prediction are missed because of many factors influence the COVID-19, e.g. geographic condition, socio-economic, government policy, etc. To handle this problem, the scenario-based prediction is proposed in this study to predict COVID-19 spread in Indonesia. This study proposed two methods to be used, i.e. Support Vector Regression (SVR) and Susceptible-Infectious-Recovered (SIR) Model. The prediction run for best-case scenario and worst-case scenario. Whereas best-case scenario used current daily case as a maximum case, worst-case scenario used another country's maximum case, i.e. India. SVR regression showed different end of epidemic, whereas best-case scenario on 21 January 2021, the worst-case scenario on 5 March 2021. SIR-Model showed the similar end of epidemic on January 2021 for both scenarios but showed the dramatically increase of infectious people from 450,000 people in best-case scenario to 5,500,000 people in worst-case scenario. The prediction can be used as an insight for the policy maker in combating the COVID-19 pandemic.