ABSTRACT
A different approach to estimation and inference with GAMs is based on represent-
ing GAMs as mixed models with the smooth terms as random effects. To facilitate
the explanation of this approach, this chapter first introduces linear mixed models,
starting with simple mixed models for balanced experimental data and then moving
on to general linear mixed models. Note that Pinheiro and Bates (2000) offers fuller
coverage of linear mixed modelling in R, while Ruppert et al. (2003) includes a clear explanation of smoothers as mixed model components.
