A0229
Title: Bayesian structure selection for vector autoregression models
Authors: Chi-Hsiang Chu - Tunghai University (Taiwan)
Mong-Na Lo - National Sun Yat-sen University (Taiwan)
Shih-Feng Huang - National Central University (Taiwan) [presenting]
Ray-Bing Chen - National Cheng Kung University (Taiwan)
Abstract: Vector autoregression (VAR) models are powerful in economic data analysis, because they can include several different time series data simultaneously. A weakness of VAR models is related to the huge coefficient dimensionality. In order to reduce the coefficient dimensionality, a Bayesian structure selection is adopted. For different types of VAR structures, different Bayesian selection approaches are proposed. The results of simulations and a real example show the performance of the proposed Bayesian approaches.
