GamFactory: An R package for integrating covariate transformations in GAMLSS

Claudia Collarin, Ca’ Foscari University of Venice

Co-authors: Matteo Fasiolo, University of Bristol

Abstract: In the Generalised Additive Models for Location, Scale and Shape (GAMLSS) framework, parsimony is essential to maintain a model that is simple, interpretable, and computationally feasible. In this work, we introduce gamFactory, an R package designed to fit GAMLSS while simultaneously transforming the covariates. A case study demonstrates the package’s application in analysing the relationship between air pollution and mortality rates in Chicago.