spikeSlabGAM: Bayesian Variable Selection, Model Choice and Regularization for Generalized Additive Mixed Models in R

TL;DR

spikeSlabGAM is an R package that applies Bayesian methods for variable selection, model choice, and regularization in generalized additive mixed models, accommodating various response types and functional forms.

Contribution

It introduces a novel spike-and-slab prior for effective selection and regularization of covariates and their functional forms in (geo-)additive mixed models.

Findings

Enables automatic covariate and interaction selection.

Determines whether effects are linear or nonlinear.

Provides regularized estimation of effect shapes.

Abstract

The R package spikeSlabGAM implements Bayesian variable selection, model choice, and regularized estimation in (geo-)additive mixed models for Gaussian, binomial, and Poisson responses. Its purpose is to (1) choose an appropriate subset of potential covariates and their interactions, (2) to determine whether linear or more flexible functional forms are required to model the effects of the respective covariates, and (3) to estimate their shapes. Selection and regularization of the model terms is based on a novel spike-and-slab-type prior on coefficient groups associated with parametric and semi-parametric effects.