Short Term Forecasting Method: Covid 19 and Capital Market in Indonesia

Abstract

This study aimed to predict the short-term confirmed Covid 19 and Jakarta Composite Index (JCI) cases in Indonesia. The prediction uses ARIMA and SutteARIMA methods, and the data processed with R software. Researcher using time series data from April 2nd, 2020 (the date of covid 19 detected in Indonesia) to September 30th, 2020. We are fitted the data with the data from October 1st to October 10th, 2020. Based on the fitted data, we could forecast the cases from October 11th to October 31st, 2020. We applied the Mean Absolute Percentage Error (MAPE) to predict accuracy measures to evaluate forecasting methods. Based on forecasting with ARIMA and SutteARIMA methods, the SutteARIMA method is more suitable than ARIMA to calculate the daily forecasts of negative Covid 19 in Indonesia with MAPE value of 0.156 (smaller than 0.21 compared to MAPE value of ARIMA). At the same time, the ARIMA method is more suitable than SutteARIMA to calculate the daily forecasts of positive Covid 19 and JCI in Indonesia. The MAPE value of 0.06 (smaller than 0.104 compared to MAPE value of SutteARIMA for positive Covid 19) and MAPE value of 0.012 (smaller than 0.021 compared to MAPE value of SutteARIMA for JCI Indonesia).

Keywords

• covid 19
• jci
• arima
• suttearima