Skip to content

Research at St Andrews

Using hierarchical centering to facilitate a reversible jump MCMC algorithm for random effects models

Research output: Contribution to journalArticle

DOI

Open Access permissions

Open

Abstract

Hierarchical centering has been described as a reparameterisation method applicable to random effects models. It has been shown to improve mixing of models in the context of Markov chain Monte Carlo (MCMC) methods. A hierarchical centering approach is proposed for reversible jump MCMC (RJMCMC) chains which builds upon the hierarchical centering methods for MCMC chains and uses them to reparameterize models in an RJMCMC algorithm. Although these methods may be applicable to models with other error distributions, the case is described for a log-linear Poisson model where the expected value λλ includes fixed effect covariates and a random effect for which normality is assumed with a zero-mean and unknown standard deviation. For the proposed RJMCMC algorithm including hierarchical centering, the models are reparameterized by modelling the mean of the random effect coefficients as a function of the intercept of the λλ model and one or more of the available fixed effect covariates depending on the model. The method is appropriate when fixed-effect covariates are constant within random effect groups. This has an effect on the dynamics of the RJMCMC algorithm and improves model mixing. The methods are applied to a case study of point transects of indigo buntings where, without hierarchical centering, the RJMCMC algorithm had poor mixing and the estimated posterior distribution depended on the starting model. With hierarchical centering on the other hand, the chain moved freely over model and parameter space. These results are confirmed with a simulation study. Hence, the proposed methods should be considered as a regular strategy for implementing models with random effects in RJMCMC algorithms; they facilitate convergence of these algorithms and help avoid false inference on model parameters.
Close

Details

Original languageEnglish
Pages (from-to)79-90
JournalComputational Statistics and Data Analysis
Volume98
Early online date2 Jan 2016
DOIs
StatePublished - Jun 2016

    Research areas

  • Combined likelihood, "Metropolis Hastings", Point transect sampling, Random effects, Reparameterisation

Discover related content
Find related publications, people, projects and more using interactive charts.

View graph of relations

Related by author

  1. Bayesian methods for hierarchical distance sampling models

    Oedekoven, C. S., Buckland, S. T., MacKenzie, M. L., King, R., Evans, K. & Burger, W. 1 Jun 2014 In : Journal of Agricultural, Biological and Environmental Statistics. 19, p. 219-239

    Research output: Contribution to journalArticle

  2. Improving distance sampling: accounting for covariates and non-independency between sampled sites

    Oedekoven, C. S., Buckland, S. T., MacKenzie, M. L., Evans, K. O. & Burger, L. W. 2013 In : Journal of Applied Ecology. 50, 3, p. 786-793 8 p.

    Research output: Contribution to journalArticle

  3. Attributing changes in the distribution of species abundance to weather variables using the example of British breeding birds

    Oedekoven, C. S., Elston, D. A., Harrison, P. J., Brewer, M. J., Buckland, S. T., Johnston, A., Foster, S. & Pearce-Higgins, J. W. 15 Jun 2017 In : Methods in Ecology and Evolution. Early View

    Research output: Contribution to journalArticle

  4. Low tortoise abundances in pine forest plantations in forest-shrubland transition areas

    Rodríguez-Caro, R. C., Oedekoven, C. S., Graciá, E., Anadón, J. D., Buckland, S. T., Esteve-Selma, M. A., Martinez, J. & Giménez, A. 8 Mar 2017 In : PLoS One. 12, 3, 13 p., e0173485

    Research output: Contribution to journalArticle

Related by journal

  1. Automatic generation of generalised regular factorial designs

    Kobilinsky, A., Monod, H. & Bailey, R. A. Sep 2017 In : Computational Statistics and Data Analysis. 113, p. 311-329 19 p.

    Research output: Contribution to journalArticle

  2. Hidden Markov models with arbitrary state dwell-time distributions

    Langrock, R. & Zucchini, W. 2011 In : Computational Statistics and Data Analysis. 55, 1, p. 715-724

    Research output: Contribution to journalArticle

  3. Default Bayesian model determination methods for generalised linear mixed models

    Overstall, A. & Forster, J. 1 Dec 2010 In : Computational Statistics and Data Analysis. 54, 12, p. 3269-3288 20 p.

    Research output: Contribution to journalArticle

  4. Estimation in Hazard Regression Models under Ordered Departures from Proportionality

    Bhattacharjee, A. Oct 2004 In : Computational Statistics and Data Analysis. 47, 3, p. 517-536 20 p.

    Research output: Contribution to journalArticle

Related by journal

  1. Computational Statistics and Data Analysis (Journal)

    Borchers, D. L. (Editor)
    2006 → …

    Activity: Publication peer-review and editorial workEditor of research journal

ID: 240103609