| Abstrakt: |
The main purpose of this article is to explore the evidence that the Bayesian Vector Autoregressive Model (BVAR) is a good forecasting model in predicting the value of the JCI index, the USD-IDR exchange rate, and the EUR-IDR exchange rate during the COVID-19 pandemic which can be seen from the forecasting accuracy value. We can see the accuracy of forecasting from the value of Mean Error (ME) to see the unfamiliarity of forecasting, Root Mean Square Error (RMSE) to see the diversity of forecasting. The BVAR model uses the Bayesian method in estimating the parameters of the vector autoregressive (VAR) model. This article uses Minnesota priors in assessing the BVAR model. The variables used are Y1 for the JCI, Y2 for the USD-IDR exchange rate, and Y3 for the EUR-IDR exchange rate. The methodology carried out, 1) Presenting descriptive statistical data in the form of boxplots, 2) seeing the stationarity of the data, looking at the ACF, PACF and ADF diagrams, if there are variables that are not stationary then differencing will be carried out, 3) Identifying the BVAR model, 4) Seeing the accuracy of forecasting BVAR against the JCI index, USD-IDR exchange rate and EUR-IDR exchange rate with ME, and RMSE indicators. The BVAR model is proven to be a good forecasting model (unbiased and small variance) in predicting the value of the JCI index, USD-IDR exchange rate, and EUR-IDR exchange rate with an ME value of 7.79007 and RMSE of 52.70694. [ABSTRACT FROM AUTHOR] |
