Updating Schemes, Correlation Structure, Blocking and Parameterization for the Gibbs Sampler
Journal of the Royal Statistical Society Series B (Statistical Methodology)Published 1 July 1997Open access
Gareth O. Roberts, Sujit K. Sahu
Citations414
SJR quartileQ1
SJR score3.31
SNIP2.48
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
Exact computable rates of convergence for Gaussian target distributions are obtained and different random and non‐random updating strategies and blocking combinations are compared using the rates.
Abstract
Summary In this paper many convergence issues concerning the implementation of the Gibbs sampler are investigated. Exact computable rates of convergence for Gaussian target distributions are obtained. Different random and non-random updating strategies and blocking combinations are compared using the rates. The effect of dimensionality and correlation structure on the convergence rates are studied. Some examples are considered to demonstrate the results. For a Gaussian image analysis problem several updating strategies are described and compared. For problems in Bayesian linear models several possible parameterizations are analysed in terms of their convergence rates characterizing the optimal choice.
Keywords
Computer ScienceMathematics
Matrix Analysis
22,234 Citations1985Roger A. Horn, Charles R. Johnson
Springer series in computational mathematicsMatrix Iterative Analysis
5,787 Citations2000Richard S. Varga
The Annals of StatisticsMarkov Chains for Exploring Posterior Distributions
3,472 Citations1994Luke Tierney
This paper outlines some of the basic methods and strategies of Markov chain methods and discusses some related theoretical and practical issues.
Journal of the American Statistical AssociationMarkov Chain Monte Carlo Convergence Diagnostics: A Comparative Review
1,977 Citations1996Mary Kathryn Cowles, Bradley P. Carlin
Journal of the Royal Statistical Society Series B (Statistical Methodology)Bayesian Computation Via the Gibbs Sampler and Related Markov Chain Monte Carlo Methods
1,727 Citations1993A. F. M. Smith, Gareth O. Roberts
The use of the Gibbs sampler for Bayesian computation and other Markov chain Monte Carlo simulation methods are reviewed and illustrated in the context of some canonical examples.
Statistical ScienceBayesian Computation and Stochastic Systems
953 Citations1995Julian Besag, Peter Green +2 more
Basic methodology of MCMC is presented, emphasizing the Bayesian paradigm, conditional probability and the intimate relationship with Markov random fields in spatial statistics, and particular emphasis on the calculation of posterior probabilities.
How Many Iterations in the Gibbs Sampler?
948 Citations1991Adrian E. Raftery, Steven M. Lewis
The Annals of StatisticsRates of convergence of the Hastings and Metropolis algorithms
570 Citations1996Kerrie Mengersen, Richard L. Tweedie
Recent results in Markov chain theory are applied to Hastings and Metropolis algorithms with either independent or symmetric candidate distributions, and it is shown geometric convergence essentially occurs if and only if $pi$ has geometric tails.
BiometrikaCovariance structure of the Gibbs sampler with applications to the comparisons of estimators and augmentation schemes
540 Citations1994Jun S. Liu, Wing Hung Wong +1 more
It is proved that Rao-Blackwellization causes a one-lag delay for the autocovariances among dependent samples obtained from data augmentation, and consequently, the mixture approximation produces estimates with smaller variances than the empirical approximation.
Image Analysis, Random Fields and Dynamic Monte Carlo Methods
514 Citations1995Gerhard Winkler
The book is mainly concerned with the mathematical foundations of Bayesian image analysis and its algorithms, which amounts to the study of Markov random fields and dynamic Monte Carlo algorithms like sampling, simulated annealing and stochastic gradient algorithms.
BiometrikaGeometric convergence and central limit theorems for multidimensional Hastings and Metropolis algorithms
412 Citations1996Gareth O. Roberts
Results on geometric ergodicity of Markov chains are developed and central limit theorems for the Metropolis algorithm are developed, showing that the negative-definiteness property is not redundant.
Physical review. D. Particles, fields, gravitation, and cosmology/Physical review. D. Particles and fieldsMultigrid Monte Carlo method. Conceptual foundations
200 Citations1989Jonathan Goodman, Alan D. Sokal
A stochastic generalization of the multigrid method, called multigrids Monte Carlo (MGMC), that reduces critical slowing down in Monte Carlo computations of lattice field theories, applicable to nonlinear {sigma} models, and to lattice gauge theories with or without bosonic matter fields.
Lecture notes in statisticsMetropolis Methods, Gaussian Proposals and Antithetic Variables
116 Citations1992Peter J. Green, Xiao-liang Han
Various aspects of a class of dynamic Monte Carlo methods, that generalises the Metropolis algorithm and includes the Gibbs sampler as a special case, are investigated, which can be used to estimate expectations of marginal distributions in stochastic systems.
Suppressing Random Walks in Markov Chain Monte Carlo Using Ordered Overrelaxation
96 Citations1998Radford M. Neal
An overrelaxed Markov chain Monte Carlo algorithm based on order statistics that can be applied whenever the full conditional distributions are such that their cumulative distribution functions and inverse cumulative distribution function can be efficiently computed.
Journal of Multivariate AnalysisComparing sweep strategies for stochastic relaxation
79 Citations1991Yali Amit, Ulf Grenander
Journal of Multivariate AnalysisOn rates of convergence of stochastic relaxation for Gaussian and non-Gaussian distributions
71 Citations1991Yali Amit
Probability in the Engineering and Informational SciencesImproving Stochastic Relaxation for Gussian Random Fields
40 Citations1990Piero Barone, Arnoldo Frigessi
A parametrized class of algorithms, which includes stochastic relaxation (Gibbs sampler), is proposed and its convergence properties are established, where a suitable choice for the parameter improves the rate of convergence with respect to stochastics relaxation for special classes of covariance matrices.
Applied and Environmental MicrobiologyEfficient parametrizations for generalized linear mixed models, (with discussion).
40 Citations1996Alan E. Gelfand, Sujit K. Sahu +1 more
To explore the molecular basis of the higher thermostability of the most thermostable PGA ever characterized, homology structural modeling and amino acid composition analyses were performed and suggested that the increased number of buried ion pair networks, lower N and Q contents, excessive arginine residues, and remarkably high content of proline residues in the structure of PGA650 could contribute to its high thermostabilities.
The Annals of StatisticsConvergence properties of the Gibbs sampler for perturbations of Gaussians
31 Citations1996Yali Amit
The Annals of Applied ProbabilityCoordinate selection rules for Gibbs sampling
19 Citations1996George S. Fishman
It is shown that neighborhood structure induces idempotency for the individual coordinate transition matrices and commutativity among subsets of these matrices corresponding to several of the plans, and encourages the belief that faster convergence for the deterministic selection rule may also imply a variance of the sample mean no larger than that arising for random selection rules.