<b>glmulti</b>: An<i>R</i>Package for Easy Automated Model Selection with (Generalized) Linear Models

TLDR

The article introduces a statistical framework and the glmulti package, applying it to simulated and real data. The paper introduces glmulti, an R package for automated model selection and multi‑model inference with generalized linear models. The package constructs all unique models from a set of explanatory variables (including optional pairwise interactions), applies user‑defined restrictions, fits them with standard glm functions, and, to handle large candidate sets, employs memory‑efficient design, parallelization, and a compiled genetic‑algorithm search. The package returns the n best models ranked by criteria such as (Q)AIC, (Q)AICc, or BIC, enabling model selection and multi‑model inference, and demonstrates its utility on simulated and real data.

Abstract

We introduce <b>glmulti</b>, an <b>R</b> package for automated model selection and multi-model inference with <code>glm</code> and related functions. From a list of explanatory variables, the provided function <code>glmulti</code> builds all possible unique models involving these variables and, optionally, their pairwise interactions. Restrictions can be specified for candidate models, by excluding specific terms, enforcing marginality, or controlling model complexity. Models are fitted with standard <b>R</b> functions like <code>glm</code>. The <i>n</i> best models and their support (e.g., (Q)AIC, (Q)AICc, or BIC) are returned, allowing model selection and multi-model inference through standard <b>R</b> functions. The package is optimized for large candidate sets by avoiding memory limitation, facilitating parallelization and providing, in addition to exhaustive screening, a compiled genetic algorithm method. This article briefly presents the statistical framework and introduces the package, with applications to simulated and real data.