ABSTRACT
This chapter’s purpose is to introduce penalized splines, show how they can be represented as mixed linear models, and emphasize some features of mixed-linear-model theory as applied to penalized splines. It considers only penalized splines for estimating a one-dimensional function of a single variable, y = f (x), i.e., scatterplot smoothing. Later chapters consider additive models with multiple x variables and spatial and other non-additive models in more than one dimension.
