ABSTRACT
This chapter explains the types of linear terms within Generalized additive models for location scale and shape (GAMLSS) models, and how they can be used. It examines linear terms and interactions for factors and numerical explanatory variables, and different useful bases used for explanatory variables. The chapter is essential for understanding the different types of additive terms in GAMLSS. The gamlss implementation uses the functions fp() and bfp(), which are loosely based on the fractional polynomial function fracpoly() for S-PLUS given by G. Ambler. Fractional polynomials can be fitted within gamlss() using the additive function fp(). It takes as arguments the x variable and npoly which takes the values 1, 2 or 3. It can be used to fit the best polynomial within a specific set of possible power values. Its argument npoly determines whether one, two or three terms in the fractional polynomial will be used in the fitting.
