ABSTRACT
This chapter introduces some multivariate linear models which are useful in regression, time series, and designed experiments. It introduces the Bayesian analysis of the multivariate versions of the regression design and autoregressive models. Linear dynamic models are linear models with parameters which always change, while structural change in linear models deals with parameters which change only once. Autoregressive processes are the most used of parametric time series models. The Bayesian analysis of the vector autoregressive process is being developed by Litterman. The chapter considers the regression model with autocorrelated error terms and multiple regression variables and the autoregressive transfer function model, which is a special case of the vector autoregressive model. The transfer function autoregressive model is a special case of the vector autoregressive model and is also a special case of the transfer function model which was analyzed by Newbold.
