ABSTRACT

This chapter describes a few special topics which extend the general linear model is useful directions. Section 13.1 introduces multivariate linear models for vector-valued responses. Longitudinal models are described in section 13.2. Section 5.5.2 discusses a non-normal alternative to the GLM via an elliptically contoured linear model. Section 13.4 discusses the hierarchical Bayesian linear model framework of Lindley and Smith, while Section 13.5 presents the dynamic linear model estimated by Kalman filtering and smoothing.