Biometrika Advance Access originally published online on November 19, 2007
Biometrika 2007 94(4):787-807; doi:10.1093/biomet/asm069
Articles |
Population-Based Reversible Jump Markov Chain Monte Carlo
Department of Mathematics, Imperial College London, London SW7 2AZ, U.K. ajay.jasra{at}imperial.ac.uk
Department of Mathematics and Statistics, McGill University, Montreal H3A 2K6, Canada dstephens{at}math.mcgill.ca
Department of Statistics, University of Oxford, Oxford OX1 3TG, U.K. cholmes{at}stats.ox.ac.uk
Received for publication 1 February 2005. Revision received 1 April 2007.
We present an extension of population-based Markov chain Monte Carlo to the transdimensional case. A major challenge is that of simulating from high- and transdimensional target measures. In such cases, Markov chain Monte Carlo methods may not adequately traverse the support of the target; the simulation results will be unreliable. We develop population methods to deal with such problems, and give a result proving the uniform ergodicity of these population algorithms, under mild assumptions. This result is used to demonstrate the superiority, in terms of convergence rate, of a population transition kernel over a reversible jump sampler for a Bayesian variable selection problem. We also give an example of a population algorithm for a Bayesian multivariate mixture model with an unknown number of components. This is applied to gene expression data of 1000 data points in six dimensions and it is demonstrated that our algorithm outperforms some competing Markov chain samplers. In this example, we show how to combine the methods of parallel chains (Geyer, 1991), tempering (Geyer & Thompson, 1995), snooker algorithms (Gilks et al., 1994), constrained sampling and delayed rejection (Green & Mira, 2001).
Key Words: Bayesian variable selection • Gene expression data • Mixture model • Reversible jump Markov chain Monte Carlo • Uniform ergodicity
References
-
Andrieu C., Moulines E. On the ergodicity properties of some adaptive Markov chain Monte Carlo algorithms. Ann. Appl. Prob. (2006) 16:1462–505.[CrossRef]
Bozdech Z., Llinas M., Pullium B. L., Wong E. D., Zhu J., DeRisi J. L. The transcriptome of the intraerythrocytic developmental cycle of Plasmodium falciparum. PLoS Biol. (2003) 1:1–16.[ISI]
Brooks S. P., Giudici P., Roberts G. O. Efficient construction of reversible jump Markov chain Monte Carlo proposal distributions (with Discussion). J. R. Statist. Soc. B (2003) 65:1–37.[CrossRef]
Cappé O., Robert C. P., Rydén T. Reversible jump, birth-and-death and more general continuous time Markov chain Monte Carlo samplers. J. R. Statist. Soc. B (2003) 65:679–700.[CrossRef]
Carlin B., Chib S. Bayesian model choice via Markov chain Monte Carlo methods. J. R. Statist. Soc. B (1995) 57:473–84.
Chauveau D., Vandekerkhove P. Improving convergence of the Hastings and Metropolis algorithm with an adaptive proposal. Scand. J. Statist. (2002) 29:13–29.[CrossRef]
Del Moral P. Feynman–Kac Formulae: Genealogical and Interacting Particle Systems with Applications (2004) New York: Springer.
Del Moral P., Doucet A., Jasra A. Sequential Monte Carlo samplers. J. R. Statist. Soc. B (2006) 68:411–36.[CrossRef]
Dellaportas P., Papageorgiou I. Multivariate mixtures of normals with an unknown number of components. Statist. Comp. (2006) 16:57–68.[CrossRef]
Denison D. G. T., Holmes C. C., Mallick B. K., Smith A. F. M. Bayesian Methods for Nonlinear Classification and Regression (2002) Chichester: Wiley.
Diaconis P., Saloff-Coste L. Comparison theorems for reversible Markov chains. Ann. Appl. Prob. (1993) 3:696–750.[CrossRef]
Dobrushin R. L. Central limit theorem for non-stationary Markov chains I. Theory Prob. Appl. (1956) 1:65–80.[CrossRef]
Keramigas E. Markov chain maximum likelihood. In: Computing Science and Statistics: The 23rd Symposium on the Interface (1991) Fairfax: Interface Foundation. 156–63.
Geyer C., Thompson E. A. Annealing Markov chain Monte Carlo with applications to ancestral inference. J. Am. Statist. Assoc. (1995) 90:909–20.[CrossRef][ISI]
Gilks W. R., Roberts G. O., George E. I. Adaptive direction sampling. Statistician (1994) 43:179–89.[CrossRef][ISI]
Godsill S. On the relationship between Markov chain Monte Carlo methods for uncertainty. J. Comp. Graph. Statist. (2001) 10:230–48.[CrossRef]
Goswami G., Liu J. S. On learning strategies for evolutionary Monte Carlo. Statist. Comp. (2007) 17:23–38.[CrossRef]
Green P. J. Reversible jump Markov chain Monte Carlo computation and Bayesian model determination. Biometrika (1995) 82:711–32.
Green P. J. Discussion of a paper by Brooks et al. J. R. Statist. Soc. B (2003) 65:48–9.
Green P. J. Highly Structured Stochastic Systems—Green P. J., Hjort N. L., Richardson S., eds. (2003) Oxford: Oxford University Press. 179–96. Trans-dimensional Markov chain Monte Carlo.
Green P. J., Mira A. Delayed rejection in reversible jump Metropolis–Hastings. Biometrika (2001) 88:1035–53.[ISI]
Heard N. A., Holmes C. C., Stephens D. A. A quantitative study of gene regulation involved in the immune response of anopheline mosquitoes: an application of Bayesian hierarchical clustering of curves. J. Am. Statist. Assoc. (2006) 101:18–29.[CrossRef][ISI]
Hodgson M. E. A. A Bayesian restoration of an ion channel signal. J. R. Statist. Soc. B (1999) 61:95–114.[CrossRef]
Hukushima K., Nemoto K. Exchange Monte Carlo method and application to spin glass simulations. J. Phys. Soc. Jap. (1996) 65:1604–8.[CrossRef]
Iba Y. Extended ensemble Monte Carlo. Int. J. Mod. Phys. (2001) 12:653–6.
Jasra A., Holmes C. C., Stephens D. A. Markov chain Monte Carlo methods and the label switching problem in Bayesian mixture modelling. Statist. Sci. (2005) 20:50–67.[CrossRef]
Jasra A., Stephens D. A., Holmes C. C. On population-based simulation. Statist. Comp. (2007) 17:263–79.[CrossRef]
Jennison C., Hurn M. A., Al-Awadhi F. Discussion of a paper by Brooks et al. J. R. Statist. Soc. B (2003) 65:44–5.
Kou S. C., Zhou Q., Wong W. H. Equi-energy sampler with applications in statistical inference and statistical mechanics. Ann. Statist. (2006) 34:1581–619.[CrossRef]
Liang F., Wong W. H. Real parameter evolutionary Monte Carlo with applications to Bayesian mixture models. J. Am. Statist. Assoc. (2001) 96:653–66.[CrossRef][ISI]
Liu J. S. Monte Carlo Strategies in Scientific Computing (2001) New York: Springer.
Madras N., Zheng Z. On the swapping algorithm. Random Struct. Algor. (2003) 22:66–97.[CrossRef]
McLachlan G. J., Peel D. Finite Mixture Models (2000) Chichester: Wiley.
Mengersen K. L., Tweedie R. L. Rates of convergence of the Hastings and Metropolis algorithms. Ann. Statist. (1996) 24:101–21.[CrossRef]
Richardson S., Green P. J. On Bayesian analysis of mixture models with an unknown number of components (with Discussion). J. R. Statist. Soc. B (1997) 59:731–92.[CrossRef]
Robert C. P., Casella G. Monte Carlo Statistical Methods (2004) 2nd. New York: Springer.
Roberts G. O., Rosenthal J. S. General state space Markov chains and Markov chain Monte Carlo algorithms. Prob. Surveys (2004) 1:20–71.[CrossRef]
Stephens M. Bayesian analysis of mixture models with an unknown number of components—an alternative to reversible jump methods. Ann. Statist. (2000) 28:40–74.[CrossRef]
Tierney L. Markov chains for exploring posterior distributions (with Discussion). Ann. Statist. (1994) 22:1701–62.[CrossRef]
Zheng Z. On swapping and simulated tempering algorithms. Stoch. Proces. Appl. (2003) 104:131–53.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||