COVID-19 Spread Pattern Using Support Vector Regression

Abstract

Pandemics are rare and happen in about 100 years period. Current pandemic, COVID-19, occurs in the industrial 4.0 era where there is a rapid development computation. Yet, the scientists in every country face difficulty in predicting the growth simulation of this pandemic. The paper tries to use a soft computing algorithm to predict the pattern of the COVID-19 pandemic in Indonesia. Support Vector Regression was used in Google Interactive Notebook with some kernels for comparison, i.e. radial basis function, linear and polynomial. The testing results showed that radial basis function outperformed other kernels as a regressor with some parameters should follows the real condition, i.e. gamma, c, and epsilon.