On the ergodicity properties of some adaptive MCMC algorithms

DOI10.1214/105051606000000286MaRDI QIDQ862214zbMATH OpenOpenAlexWikidataFDO

Authors Christophe Andrieu, Eric Moulines

Publication date 5 February 2007

Published in The Annals of Applied Probability (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/math/0610317

zbMATH Keywords

Metropolis-Hastings algorithmmartingaleself-tuning algorithmStochastic approximationAdaptive Markov chain Monte Carlo algorithmPoisson methodrandomly varying truncationstate-dependent noise

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) Continuous-time Markov processes on discrete state spaces (60J27) Transition functions, generators and resolvents (60J35)

Abstract: In this paper we study the ergodicity properties of some adaptive Markov chain Monte Carlo algorithms (MCMC) that have been recently proposed in the literature. We prove that under a set of verifiable conditions, ergodic averages calculated from the output of a so-called adaptive MCMC sampler converge to the required value and can even, under more stringent assumptions, satisfy a central limit theorem. We prove that the conditions required are satisfied for the independent Metropolis--Hastings algorithm and the random walk Metropolis algorithm with symmetric increments. Finally, we propose an application of these results to the case where the proposal distribution of the Metropolis--Hastings update is a mixture of distributions from a curved exponential family.

Recommendations

• On the stability and ergodicity of adaptive scaling Metropolis algorithms
• Limit theorems for some adaptive MCMC algorithms with subgeometric kernels
• Adaptive independent Metropolis-Hastings
• On the efficiency of adaptive MCMC algorithms
• On adaptive Markov chain Monte Carlo algorithms

Cites work

• scientific article; zbMATH DE number 3723610 (Why is no real title available?)
• scientific article; zbMATH DE number 48727 (Why is no real title available?)
• scientific article; zbMATH DE number 976356 (Why is no real title available?)
• scientific article; zbMATH DE number 1043533 (Why is no real title available?)
• scientific article; zbMATH DE number 1086057 (Why is no real title available?)
• scientific article; zbMATH DE number 3992716 (Why is no real title available?)
• scientific article; zbMATH DE number 3405266 (Why is no real title available?)
• A Liapounov bound for solutions of the Poisson equation
• A maximal inequality and dependent strong laws
• An adaptive Metropolis algorithm
• Computable bounds for geometric convergence rates of Markov chains
• Convergence and robustness of the Robbins-Monro algorithm truncated at randomly varying bounds
• Finite Mixture Distributions, Sequential Likelihood and the EM Algorithm
• Geometric convergence and central limit theorems for multidimensional Hastings and Metropolis algorithms
• Geometric ergodicity of Metropolis algorithms
• Inference from iterative simulation using multiple sequences
• Markov chains and stochastic stability
• On adaptive Markov chain Monte Carlo algorithms
• On an adaptive version of the Metropolis–Hastings algorithm with independent proposal distribution
• On the Poisson equation in the potential theory of a single kernel.
• On the convergence properties of the EM algorithm
• On the ergodicity properties of some adaptive MCMC algorithms
• Rates of convergence of the Hastings and Metropolis algorithms
• Renewal theory and computable convergence rates for geometrically erdgodic Markov chains
• Some Multivariate Chebyshev Inequalities with Extensions to Continuous Parameter Processes
• Stability of Stochastic Approximation under Verifiable Conditions
• Statistical analysis of finite mixture distributions

Cited in

(88)

• Transport map accelerated Markov chain Monte Carlo
• Efficient computational strategies for doubly intractable problems with applications to Bayesian social networks
• Generalized darting Monte Carlo
• A posteriori stochastic correction of reduced models in delayed-acceptance MCMC, with application to multiphase subsurface inverse problems
• Adaptive tempered reversible jump algorithm for Bayesian curve fitting
• Computation for latent variable model estimation: a unified stochastic proximal framework
• On‐line partitioning of the sample space in the regional adaptive algorithm
• A central limit theorem for adaptive and interacting Markov chains
• Adaptive Incremental Mixture Markov Chain Monte Carlo
• Posterior inference on parameters of stochastic differential equations via non-linear Gaussian filtering and adaptive MCMC
• Consistency of Markov chain quasi-Monte Carlo on continuous state spaces
• On the stability of some controlled Markov chains and its applications to stochastic approximation with Markovian dynamic
• Adaptive Metropolis-Hastings sampling using reversible dependent mixture proposals
• Weak Convergence Rates of Population Versus Single-Chain Stochastic Approximation MCMC Algorithms
• Integrating production data under uncertainty by parallel interacting Markov chains on a reduced dimensional space
• Bayesian inversion by parallel interacting Markov chains
• Adaptive Metropolis algorithm using variational Bayesian adaptive Kalman filter
• Fluctuations of interacting Markov chain Monte Carlo methods
• Algorithms for Kullback-Leibler approximation of probability measures in infinite dimensions
• Sampling constrained probability distributions using spherical augmentation
• On the stability of sequential Monte Carlo methods in high dimensions
• Kernel estimators of asymptotic variance for adaptive Markov chain Monte Carlo
• A note on formal constructions of sequential conditional couplings
• Sequentially interacting Markov chain Monte Carlo methods
• An adaptive multiple-try Metropolis algorithm
• Relative entropy minimization over Hilbert spaces via Robbins-Monro
• On the convergence of adaptive sequential Monte Carlo methods
• Grapham: graphical models with adaptive random walk Metropolis algorithms
• A cautionary tale on the efficiency of some adaptive Monte Carlo schemes
• Limit theorems for some adaptive MCMC algorithms with subgeometric kernels. II
• Dimension-independent likelihood-informed MCMC
• Bayesian computation: a summary of the current state, and samples backwards and forwards
• Automatically tuned general-purpose MCMC via new adaptive diagnostics
• An overview of stochastic approximation Monte Carlo
• A non linear mixed effects model of plant growth and estimation via stochastic variants of the EM algorithm
• Convergence of adaptive and interacting Markov chain Monte Carlo algorithms
• Adaptive Gibbs samplers and related MCMC methods
• Trajectory averaging for stochastic approximation MCMC algorithms
• Adaptive independent Metropolis-Hastings
• Preconditioned Bayesian regression for stochastic chemical kinetics
• Accelerated dimension-independent adaptive metropolis
• On adaptive Metropolis-Hastings methods
• Construction of Bayesian deformable models via a stochastic approximation algorithm: a convergence study
• Inference for Lévy-driven stochastic volatility models via adaptive sequential Monte Carlo
• Estimating discrete Markov models from various incomplete data schemes
• Robust adaptive Metropolis algorithm with coerced acceptance rate
• Calibration and filtering for multi factor commodity models with seasonality: incorporating panel data from futures contracts
• Stability of adversarial Markov chains, with an application to adaptive MCMC algorithms
• Adaptive sequential Monte Carlo by means of mixture of experts
• Efficient MCMC sampling in dynamic mixture models
• Multilevel dimension-independent likelihood-informed MCMC for large-scale inverse problems
• Convergence of Markovian stochastic approximation with discontinuous dynamics
• Skew mixture models for loss distributions: a Bayesian approach
• Bayesian quantile regression using the skew exponential power distribution
• Convergence rates of Metropolis-Hastings algorithms
• Importance sampling correction versus standard averages of reversible MCMCs in terms of the asymptotic variance
• Conditional particle filters with diffuse initial distributions
• Efficient stochastic optimisation by unadjusted Langevin Monte Carlo. Application to maximum marginal likelihood and empirical Bayesian estimation
• Unified Bayesian conditional autoregressive risk measures using the skew exponential power distribution
• A strong law of large numbers for strongly mixing processes
• An adaptive version for the Metropolis adjusted Langevin algorithm with a truncated drift
• On nonlinear Markov chain Monte Carlo
• Ergodicity of combocontinuous adaptive MCMC algorithms
• AMCMC: an R interface for adaptive MCMC
• Markov chain Monte Carlo algorithms with sequential proposals
• A framework for adaptive MCMC targeting multimodal distributions
• Markovian stochastic approximation with expanding projections
• f-SAEM: a fast stochastic approximation of the EM algorithm for nonlinear mixed effects models
• Ensemble transport adaptive importance sampling
• Interacting sequential Monte Carlo samplers for trans-dimensional simulation
• Geometric adaptive Monte Carlo in random environment
• Bayesian nonparametric vector autoregressive models
• Advanced Multilevel Monte Carlo Methods
• Stochastic proximal-gradient algorithms for penalized mixed models
• On the efficiency of adaptive MCMC algorithms
• Adaptive Markov chain Monte Carlo for auxiliary variable method and its application to parallel tempering
• Coupling and Ergodicity of Adaptive Markov Chain Monte Carlo Algorithms
• Limit theorems for some adaptive MCMC algorithms with subgeometric kernels
• On the ergodicity of the adaptive Metropolis algorithm on unbounded domains
• Convergent stochastic expectation maximization algorithm with efficient sampling in high dimension. Application to deformable template model estimation
• A multi-point Metropolis scheme with generic weight functions
• On the stability and ergodicity of adaptive scaling Metropolis algorithms
• Parametric inference for mixed models defined by stochastic differential equations
• Simple Monte Carlo and the Metropolis algorithm
• On the ergodicity properties of some adaptive MCMC algorithms
• Bayesian inversion for anisotropic hydraulic phase-field fracture
• Adaptive Component-Wise Multiple-Try Metropolis Sampling
• An adaptive approach to Langevin MCMC

This page was built for publication: On the ergodicity properties of some adaptive MCMC algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q862214)