Extending integrated nested laplace approximation to a class of near-gaussian latent models

Thiago G. Martins*, Håvard Rue

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

8 Scopus citations

Abstract

This work extends the integrated nested Laplace approximation (INLA) method to latent models outside the scope of latent Gaussian models, where independent components of the latent field can have a near-Gaussian distribution. The proposed methodology is an essential component of a bigger project that aims to extend the R package INLA in order to allow the user to add flexibility and challenge the Gaussian assumptions of some of the model components in a straightforward and intuitive way. Our approach is applied to two examples, and the results are compared with that obtained by Markov chain Monte Carlo, showing similar accuracy with only a small fraction of computational time. Implementation of the proposed extension is available in the R-INLA package.

Original languageEnglish (US)
Pages (from-to)893-912
Number of pages20
JournalScandinavian Journal of Statistics
Volume41
Issue number4
DOIs
StatePublished - Dec 1 2014
Externally publishedYes

Keywords

  • Approximate bayesian inference
  • Integrated nested laplace approximation
  • Markov chain monte carlo
  • Near-gaussian latent models

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Fingerprint

Dive into the research topics of 'Extending integrated nested laplace approximation to a class of near-gaussian latent models'. Together they form a unique fingerprint.

Cite this