A Metropolis-Hastings based method for sampling from the \(G\)-Wishart distribution in Gaussian graphical models. (English) Zbl 1274.65009

Summary: In Gaussian graphical models, the conjugate prior for the precision matrix \(K\) is called \(G\)-Wishart distribution, \(W_{G}(\delta ,D)\). In this paper we propose a new sampling method for the \(W_{G}(\delta,D)\) based on the Metropolis-Hastings algorithm and we show its validity through a number of numerical experiments. We show that this method can be easily used to estimate the Deviance Information Criterion, providing with a computationally inexpensive approach for model selection.


65C60 Computational problems in statistics (MSC2010)


Full Text: DOI Euclid


[1] Asci, C. and Piccioni, M. (2007). Functionally compatible local characteristics for the local specification of priors in graphical models., Scandinavian Journal of Statistics 34 829-840. · Zbl 1164.62004
[2] Atay-Kayis, A. and Massam, H. (2005). A Monte Carlo method for computing the marginal likelihood in nondecomposable Gaussian graphical models., Biomertika 92 317-335. · Zbl 1094.62028 · doi:10.1093/biomet/92.2.317
[3] Carvalho, C. M., Massam, H. and West, M. (2007). Simulation of Hyper Inverse-Wishart Distributions in Graphical Models., Biometrika 94 647-659. · Zbl 1135.62011 · doi:10.1093/biomet/asm056
[4] Diaconis, P. and Ylvisaker, D. (1979). Conjugate priors for exponential families., Annals of Statistics 7 269-281. · Zbl 0405.62011 · doi:10.1214/aos/1176344611
[5] Fisher, R. A. (1936). The Use of Multiple Measurements in Taxonomy Problems., Annals of Eugenics 7 179-188.
[6] Gelman, A., Carlin, J. B., Stern, H. S. and Rubin, D. B. (1995)., Bayesian data analysis . Chapman and Hall. · Zbl 1279.62004
[7] Hastie, T., Tibshirani, R. and Freedman, J. (2009)., Elements of Statistical Learning , Second ed. Springer-Verlag. · Zbl 1273.62005
[8] Lauritzen, L. S. (1996)., Graphical Models . Oxford University Press. · Zbl 0907.62001
[9] Lenkoski, A. and Dobra, A. (2010). Computational Aspects Related to Inference in Gaussian Graphical Models with the G-Wishart Prior., Journal of Computational and Graphical Statistics .
[10] McHugh, J. (1990)., Algorithmic Graph Theory . Prentice-Hall. · Zbl 0755.68056
[11] Mengersen, K. L. and Tweedie, R. L. (1996). Rates of Convergence of the Hastings and Metropolis Algorithms., The Annals of Statistics 24 101-121. · Zbl 0854.60065 · doi:10.1214/aos/1033066201
[12] Piccioni, M. (2000). Independence structure of natural conjugate densities to exponential families and the Gibbs Sampler., Scandinavian Journal of Statistics 27 111-127. · Zbl 0938.62025 · doi:10.1111/1467-9469.00182
[13] Roverato, A. (2000). Cholesky decomposition of a hyper inverse Wishart matrix., Biometrika 87 99-112. · Zbl 0974.62047 · doi:10.1093/biomet/87.1.99
[14] Roverato, A. (2002). Hyper Inverse Wishart Distribution for Non-decomposable Graphs and its Application to Bayesian Inference for Gaussian Graphical Models., Scandinavian Journal of Statistics 29 391-411. · Zbl 1036.62027 · doi:10.1111/1467-9469.00297
[15] Spiegelhalter, D. J., Best, N. G., Carlin, B. P. and van der Linde, A. (2002). Bayesian measures of model complexity and fit (with discussion)., Journal of the Royal Statistical Society, Series B (Statistical Methodology) 64 583-639. · Zbl 1067.62010 · doi:10.1111/1467-9868.00353
[16] Tierney, L. (1994). Markov Chains for Exploring Posterior Distributions., The Annals of Statistics 22 1701-1762. · Zbl 0829.62080 · doi:10.1214/aos/1176325750
[17] Wang, H. and Carvalho, C. M. (2010). Simulation of Hyper-Inverse Wishart Distributions for Non-decomposable Graphs., Electronic Journal of Statistics 4 1470-1475. · Zbl 1329.60008 · doi:10.1214/10-EJS591
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.