Measuring the Accuracy of Forecasting Using Integrated VAR models

Abstract

This study addressed the problem of prediction integration. Different weighting methods are applied to different VAR models. Some economic time series such as unemployment rates, economic growth rates and the general government expenditure series are used to study their effect on each other through the use of VAR , VARX and SVAR models. An evaluation of the integration between predections is presented. The results showed that the combined models prediction is better than normal models. The BICW method is the best combining method.