Production forecasting of Taiwan's technology industrial cluster: A Bayesian autoregression approach

Abstract

This study proposes a forecasting method that combines the clustering effect and non-informative diffiuse-prior Bayesian vector autoregression (NDBVAR) model to forecast the productions of technology industries. Two empirical cases are examined to verify the proposed method: the semiconductor industry and computer manufacturing industry in Taiwan. It is found that the NDBVAR model outperforms the other three conventional time series models including the autoregression (AR), vector autoregression (VAR), and Litterman Bayesian VAR (LBVAR) models. Moreover the NDBVAR model also outperforms the forecast reports from leading market information providers over the past several years. The forecasting method proposed is therefore concluded to be a feasible approach for production prediction, especially for technology, industries in volatile environments.