Markov chain Monte Carlo confidence intervals

DOI10.3150/15-BEJ712zbMATH Open1345.60074arXiv1209.0703MaRDI QIDQ282567FDOQ282567

Authors: Yves F. Atchadé 

Publication date: 12 May 2016

Published in: Bernoulli (Search for Journal in Brave)

Abstract: For a reversible and ergodic Markov chain $X_n,ngeq0$ with invariant distribution $pi$, we show that a valid confidence interval for $pi(h)$ can be constructed whenever the asymptotic variance $sigm{a}_P^2(h)$ is finite and positive. We do not impose any additional condition on the convergence rate of the Markov chain. The confidence interval is derived using the so-called fixed-b lag-window estimator of $sigm{a}_P^2(h)$. We also derive a result that suggests that the proposed confidence interval procedure converges faster than classical confidence interval procedures based on the Gaussian distribution and standard central limit theorems for Markov chains.

Full work available at URL: https://arxiv.org/abs/1209.0703

Recommendations

• On confidence intervals from simulation of finite Markov chains
• Estimation of the Asymptotic Variance in the CLT for Markov Chains
• How to combine fast heuristic Markov chain Monte Carlo with slow exact sampling
• On the validity of the batch quantile method for Markov chains
• Nonasymptotic bounds on the estimation error of MCMC algorithms

zbMATH Keywords

confidence intervalsMarkov chainsBerry-Esseen boundsMonte Carlo methodmartingale approximationlag-window estimators

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10)

Cites Work

• Exponential convergence of Langevin distributions and their discrete approximations
• Markov chains for exploring posterior distributions. (With discussion)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Stochastic simulation: Algorithms and analysis
• Markov chains and stochastic stability
• General state space Markov chains and MCMC algorithms
• Central limit theorem for additive functionals of reversible Markov processes and applications to simple exclusions
• Polynomial convergence rates of Markov chains
• Batch means and spectral variance estimators in Markov chain Monte Carlo
• Fixed-Width Output Analysis for Markov Chain Monte Carlo
• Optimal Bandwidth Selection in Heteroskedasticity–Autocorrelation Robust Testing
• Automatic Lag Selection in Covariance Matrix Estimation
• Simple Robust Testing of Regression Hypotheses
• On variance conditions for Markov chain CLTs
• An adaptive version for the Metropolis adjusted Langevin algorithm with a truncated drift
• A martingale decomposition for quadratic forms of Markov chains (with applications)
• On Consistent Estimates of the Spectrum of a Stationary Time Series
• Confidence Interval Estimation Using Standardized Time Series
• A NEW ASYMPTOTIC THEORY FOR HETEROSKEDASTICITY-AUTOCORRELATION ROBUST TESTS
• Simulation Output Analysis Using Standardized Time Series
• Mean-Square Consistency of the Variance Estimator in Steady-State Simulation Output Analysis
• Title not available (Why is that?)
• The Error in Rejection Probability of Simple Autocorrelation Robust Tests
• Kernel estimators of asymptotic variance for adaptive Markov chain Monte Carlo
• Term-by-Term Differentiability of Mercer's Expansion
• An Improved Formula for the Asymptotic Variance of Spectrum Estimates
• Statistical Spectral Analysis of Time Series Arising from Stationary Stochastic Processes
• On mean central limit theorems for stationary sequences

Cited In (9)

• Confidence intervals through sequential Monte Carlo
• Title not available (Why is that?)
• Analyzing Markov chain Monte Carlo output
• Bayesian approach for confidence intervals of variance on the normal distribution
• Convergence rates of Metropolis-Hastings algorithms
• Recursive construction of confidence regions
• Adaptive confidence bands for Markov chains and diffusions: estimating the invariant measure and the drift
• Mixing time estimation in reversible Markov chains from a single sample path
• Simple confidence intervals for MCMC without CLTs

This page was built for publication: Markov chain Monte Carlo confidence intervals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q282567)