Adaptive optimal scaling of Metropolis-Hastings algorithms using the Robbins-Monro process

DOI10.1080/03610926.2014.936562MaRDI QIDQ2817137zbMATH OpenOpenAlexFDO

Authors Paul H. Garthwaite, Yanan Fan, S. A. Sisson

Publication date 29 August 2016

Published in Communications in Statistics. Theory and Methods (Search for Journal in Brave)

Copyright license Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International

Full work available at URL https://arxiv.org/abs/1006.3690

zbMATH Keywords

\(\alpha\)-mixingBerry-Esseen type boundconditional modeMetropolis-Hastings algorithmsrandom-walktruncated data

Mathematics Subject Classification ID

Sums of independent random variables; random walks (60G50) Generalized linear models (logistic models) (62J12) Numerical analysis or methods applied to Markov chains (65C40)

Abstract: We present an adaptive method for the automatic scaling of Random-Walk Metropolis-Hastings algorithms, which quickly and robustly identifies the scaling factor that yields a specified overall sampler acceptance probability. Our method relies on the use of the Robbins-Monro search process, whose performance is determined by an unknown steplength constant. We give a very simple estimator of this constant for proposal distributions that are univariate or multivariate normal, together with a sampling algorithm for automating the method. The effectiveness of the algorithm is demonstrated with both simulated and real data examples. This approach could be implemented as a useful component in more complex adaptive Markov chain Monte Carlo algorithms, or as part of automated software packages.

Recommendations

• Optimal scaling for various Metropolis-Hastings algorithms.
• On adaptive Markov chain Monte Carlo algorithms
• Optimal scaling of Metropolis algorithms: Heading toward general target distributions
• Optimal proposal distributions and adaptive MCMC
• Weak convergence and optimal scaling of random walk Metropolis algorithms

Cites work

• A Stochastic Approximation Method
• An adaptive Metropolis algorithm
• Bias correction in generalised linear mixed models with a single component of dispersion
• Componentwise adaptation for high dimensional MCMC
• Confidence Intervals from Randomization Tests
• General design Bayesian generalized linear mixed models
• Generalised linear mixed model analysis via sequential Monte Carlo sampling
• Generating Monte Carlo Confidence Intervals by the Robbins-Monro Process
• Handbook of Markov Chain Monte Carlo
• scientific article; zbMATH DE number 3535514 (Why is no real title available?)
• Learn from thy neighbor: parallel-chain and regional adaptive MCMC
• On a stochastic approximation procedure based on averaging
• On the stability and ergodicity of adaptive scaling Metropolis algorithms
• Optimal scaling for various Metropolis-Hastings algorithms.
• Prepivoting by weighted bootstrap iteration
• Semiparametric regression for clustered data
• Weak convergence and optimal scaling of random walk Metropolis algorithms

Cited in

(33)

• Robust adaptive Metropolis algorithm with coerced acceptance rate
• A sampling algorithm for bandwidth estimation in a nonparametric regression model with a flexible error density
• A rare event approach to high-dimensional approximate Bayesian computation
• Simultaneous fitting of Bayesian penalised quantile splines
• A Bayesian procedure for bandwidth selection in circular kernel density estimation
• Updating variational Bayes: fast sequential posterior inference
• Bayesian bandwidth estimation for a semi-functional partial linear regression model with unknown error density
• Scalable Bayesian inference for the inverse temperature of a hidden Potts model
• Computationally efficient Bayesian estimation of high-dimensional Archimedean copulas with discrete and mixed margins
• A Bayesian approach for determining the optimal semi-metric and bandwidth in scalar-on-function quantile regression with unknown error density and dependent functional data
• Bayesian bandwidth estimation for local linear fitting in nonparametric regression models
• Optimal proposal distributions and adaptive MCMC
• Bayesian Inference for the One-Factor Copula Model
• Efficient Bayesian Inference for Nonlinear State Space Models With Univariate Autoregressive State Equation
• Estimation of a functional single index model with dependent errors and unknown error density
• Mixing of MCMC algorithms
• Bayesian bandwidth estimation and semi-metric selection for a functional partial linear model with unknown error density
• Bayesian models for spatial count data with informative finite populations with application to the American community survey
• Accelerating adaptation in the adaptive Metropolis–Hastings random walk algorithm
• On the stability and ergodicity of adaptive scaling Metropolis algorithms
• Empirical Bayes inference for the block maxima method
• Tuning diagonal scale matrices for HMC
• Bayesian flexible local projections
• Adaptive reduced tempering for Bayesian inverse problems and rare event simulation
• Adaptive reduced multilevel splitting
• Personalized biopsy schedules using an interval-censored cause-specific joint model
• A joint model for (un)bounded longitudinal markers, competing risks, and recurrent events using patient registry data
• Locally time-varying parameter regression
• The Block-Correlated Pseudo Marginal Sampler for State Space Models
• Analysing symbolic data by pseudo-marginal methods
• Estimation and uncertainty quantification for extreme quantile regions
• On adaptive Metropolis-Hastings methods
• On adaptive Markov chain Monte Carlo algorithms

This page was built for publication: Adaptive optimal scaling of Metropolis-Hastings algorithms using the Robbins-Monro process

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2817137)