Covid-19 Cases in Malaysia: Time Series Forecasting and The Uncertainty

Abstract

COVID-19 predictive analysis has emerged as a prominent research topic to assist healthcare services and governments in planning and controlling the development of the infectious illness. Modelling and forecasting the virus's daily transmission pattern can help health-care systems prepare for the influx of new patients. In fact, an accurate forecast of COVID-19 transmission is crucial for all parties involved. Therefore, daily new cases of COVID-19 is forecast using three different methods such as ARIMA method, Holt’s Linear Trend method and Naïve method. The methods is then compared using RMSE, MAE, MAPE and the uncertainty were calculated. The ARIMA has the lowest level for all performance measurement but the Naïve method produce the lowest uncertainty. The forecast value of COVID-19 cases from 26/07/2022 to 30/07/2022 using ARIMA(2,1,4) fluctuates from 3516 cases to 4237 cases. Cases prediction that are accurate can be used as information as it could help the government, healthcare management and citizens prepare for the wave of new cases and help for a similar pandemic issues in future.