Implementation of Holt-Winter Exponential Smoothing Method to Forecast The Spread of Covid-19

Abstract

This research examines the Holt winters exponential smoothing method to solve forecasting problems using case studies of the spread of Covid-19. The data source uses the transmission of Covid-19 in Indonesia. MAPE is used to measure errors in data forecasting. The results are structured to serve as a recommendation for other researchers in choosing a method for predicting the spread of the disease. Based on the results, forecasting with the Holt-Winters model in positive cases produces a MAPE value of 9.21% using the Multiplicative model and the best parameter values of alpha, beta and gamma (0.4, 0.05, 0.25). Whereas in the case of human recovery, the MAPE value was 11.86% using the Multiplicative model and the best values for the parameters alpha, beta and gamma (0.4, 0.03, 0.1). And in the case of death it produces a MAPE of 17.97% using the Multiplicative model and the parameter values alpha, beta and gamma (0.4, 0.01, 0.1). So, it can be concluded that the Holt-Winters method on human positive case data shows a good outcome performance while recovered and deceased cases produce a well-being analysis performance.