G5 Doctoral dissertation (article)
On modeling multivariate abundance data with generalized linear latent variable models (2020)

Niku, J. (2020). On modeling multivariate abundance data with generalized linear latent variable models [Doctoral dissertation]. Jyväskylän yliopisto. JYU Dissertations, 192. http://urn.fi/URN:ISBN:978-951-39-8062-7

JYU authors or editors

Publication details

All authors or editors: Niku, Jenni

eISBN: 978-951-39-8062-7

Journal or series: JYU Dissertations

eISSN: 2489-9003

Publication year: 2020

Number in series: 192

Publisher: Jyväskylän yliopisto

Place of Publication: Jyväskylä

Publication country: Finland

Publication language: English

Persistent website address: http://urn.fi/URN:ISBN:978-951-39-8062-7

Publication open access: Openly available

Publication channel open access: Open Access channel

Abstract

The multivariate abundance data consist typically of multiple, correlated species encountered at a set of sites, together with records of additional covariates. When analysing such data, model-based approaches have been shown to outperform classical algorithmic-based dimension reduction methods. In this thesis we consider generalized linear latent variable models, which offer a general framework for the analysis of multivariate abundance data. In order to make the models more attractive among practitioners, new computationally efficient algorithms for the parameter estimation are developed by applying closed form approximation methods, the variational approximation method and the Laplace approximation method, for the marginal likelihood and by utilizing automatic differentiation tools when implementing the algorithms. The accuracy and computational efficiency of the methods are investigated and compared to existing methods through extensive simulation studies. The developed algorithms and additional tools implemented for model diagnosis, visualization and statistical inference are collected in R package gllvm. Several examples are provided to illustrate the use of the generalized linear latent variable models in ordination and when studying the between-species correlations and the effects of environmental variables, trait variables and their interactions on ecological communities.

Keywords: statistical models; multivariable methods; linear models; approximation; ecology; biotic communities; biodiversity

Free keywords: Community analysis; ecological data; fourth-corner models; generalized linear models; joint modeling; Laplace approximation; latent variables; multivariate analysis; ordination; species interactions; variational approximation

Contributing organizations

Ministry reporting: Yes

Reporting Year: 2020