Demonstration of the gllvm R-package (digital)

Bert van der Veen, Norwegian Institute of Bioeconomy Research

I was invited by Jonas Lambrechts.

Multivariate analysis is deeply rooted in various fields in ecology. Most notably, community ecology has been using methods for the dimension reduction of multivariate data for over 70 years. In recent decades, statisticians have developed new methods for ordination that lean heavily on Generalized Linear Models and on the mixed-effects modeling framework. These models are referred to as Joint Species (Abundance) Distribution Models, or alternatively as model-based ordination, as part of the Generalized Linear Latent Variable Modeling (GLLVM) framework. GLLVMs are a statistically rigorous method for ordination with dedicated developments in unconstrained (Hui et al. 2015) and constrained (van der Veen et al. in review) ordination, but allow for much more flexibility than classical ordination methods do. This presentation will give an overview of the possibilities that GLLVMs offer for multivariate analysis in ecology, and demonstrate some implementations with the gllvm R-package.

The material is available on github: https://github.com/BertvanderVeen/Antwerp2023