A hybrid approximation bayesian test of variance components for longitudinal data

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

The test of variance components of possibly correlated random effects in generalized linear mixed models (GLMMs) can be used to examine if there exists heterogeneous effects. The Bayesian test with Bayes factors offers a flexible method. In this article, we focus on the performance of Bayesian tests under three reference priors and a conjugate prior: an approximate uniform shrinkage prior, modified approximate Jeffreys' prior, half-normal unit information prior and Wishart prior. To compute Bayes factors, we propose a hybrid approximation approach combining a simulated version of Laplace's method and importance sampling techniques to test the variance components in GLMMs.

Original languageEnglish
Pages (from-to)2849-2864
Number of pages16
JournalCommunications in Statistics - Theory and Methods
Volume39
Issue number16
DOIs
Publication statusPublished - 2010 Aug 19

Fingerprint

Components of Variance
Longitudinal Data
Generalized Linear Mixed Model
Bayes Factor
Approximation
Reference Prior
Laplace's Method
Jeffreys Prior
Unit normal vector
Conjugate prior
Variance Components
Importance Sampling
Prior Information
Shrinkage
Random Effects

All Science Journal Classification (ASJC) codes

  • Statistics and Probability

Cite this

@article{d45ee8e68c164558a023e1db076c936a,
title = "A hybrid approximation bayesian test of variance components for longitudinal data",
abstract = "The test of variance components of possibly correlated random effects in generalized linear mixed models (GLMMs) can be used to examine if there exists heterogeneous effects. The Bayesian test with Bayes factors offers a flexible method. In this article, we focus on the performance of Bayesian tests under three reference priors and a conjugate prior: an approximate uniform shrinkage prior, modified approximate Jeffreys' prior, half-normal unit information prior and Wishart prior. To compute Bayes factors, we propose a hybrid approximation approach combining a simulated version of Laplace's method and importance sampling techniques to test the variance components in GLMMs.",
author = "Tsai, {Miao Yu}",
year = "2010",
month = "8",
day = "19",
doi = "10.1080/03610920903145154",
language = "English",
volume = "39",
pages = "2849--2864",
journal = "Communications in Statistics - Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "16",

}

A hybrid approximation bayesian test of variance components for longitudinal data. / Tsai, Miao Yu.

In: Communications in Statistics - Theory and Methods, Vol. 39, No. 16, 19.08.2010, p. 2849-2864.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A hybrid approximation bayesian test of variance components for longitudinal data

AU - Tsai, Miao Yu

PY - 2010/8/19

Y1 - 2010/8/19

N2 - The test of variance components of possibly correlated random effects in generalized linear mixed models (GLMMs) can be used to examine if there exists heterogeneous effects. The Bayesian test with Bayes factors offers a flexible method. In this article, we focus on the performance of Bayesian tests under three reference priors and a conjugate prior: an approximate uniform shrinkage prior, modified approximate Jeffreys' prior, half-normal unit information prior and Wishart prior. To compute Bayes factors, we propose a hybrid approximation approach combining a simulated version of Laplace's method and importance sampling techniques to test the variance components in GLMMs.

AB - The test of variance components of possibly correlated random effects in generalized linear mixed models (GLMMs) can be used to examine if there exists heterogeneous effects. The Bayesian test with Bayes factors offers a flexible method. In this article, we focus on the performance of Bayesian tests under three reference priors and a conjugate prior: an approximate uniform shrinkage prior, modified approximate Jeffreys' prior, half-normal unit information prior and Wishart prior. To compute Bayes factors, we propose a hybrid approximation approach combining a simulated version of Laplace's method and importance sampling techniques to test the variance components in GLMMs.

UR - http://www.scopus.com/inward/record.url?scp=77955576385&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955576385&partnerID=8YFLogxK

U2 - 10.1080/03610920903145154

DO - 10.1080/03610920903145154

M3 - Article

AN - SCOPUS:77955576385

VL - 39

SP - 2849

EP - 2864

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 16

ER -