ABSTRACT
This chapter provides a brief introduction to linear models which exhibit a change in some or all of the parameters during the period of observing the data. These models are quite useful for some economic, social, physical, and biological processes. Structural change in linear models is easily generalized from changing normal sequences. From the Bayesian viewpoint, D. Holbert and L. D. Broemeling studied two-phase regression by assigning a marginal uniform proper prior to the shift point and an improper prior for the unknown regression parameters. The chapter presents the posterior analysis of a changing linear model. It provides an example which illustrates the method of estimating the shift point and all unknown regression parameters of a two-phase regression model. The example demonstrates the sensitivity of the marginal posterior distribution of switch point to different prior distributions, and consists of providing inferences for all the parameters.
