Efficient shape-constrained inference for the autocovariance sequence from a reversible Markov chain

DOI10.1214/23-AOS2335MaRDI QIDQ6183871OpenAlexFDO

Authors Stephen Berg, Hyebin Song

Publication date 4 January 2024

Published in The Annals of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2207.12705

zbMATH Keywords

Markov chain Monte Carloasymptotic varianceshape-constrained inferenceautocovariance sequence estimation

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Nonparametric estimation (62G05) Discrete-time Markov processes on general state spaces (60J05)

Abstract: In this paper, we study the problem of estimating the autocovariance sequence resulting from a reversible Markov chain. A motivating application for studying this problem is the estimation of the asymptotic variance in central limit theorems for Markov chains. The asymptotic variance quantifies uncertainties in averages of the form $M^{-1}su{m}_{t=0}^{M-1}g(X_t)$, where $X_0,X_1,\mathrm{...}$ are iterates from a Markov chain. It is well known that the autocovariances from reversible Markov chains can be represented as the moments of a unique positive measure supported on $[-1,1]$. We propose a novel shape-constrained estimator of the autocovariance sequence. Our approach is based on the key observation that the representability of the autocovariance sequence as a moment sequence imposes certain shape constraints, which we can exploit in the estimation procedure. We examine the theoretical properties of the proposed estimator and provide strong consistency guarantees for our estimator. In particular, for reversible Markov chains satisfying a geometric drift condition, we show that our estimator is strongly consistent for the true autocovariance sequence with respect to an $el{l}_2$ distance, and that our estimator leads to strongly consistent estimates of the asymptotic variance. Finally, we perform empirical studies to illustrate the theoretical properties of the proposed estimator as well as to demonstrate the effectiveness of our estimator in comparison with other current state-of-the-art methods for Markov chain Monte Carlo variance estimation, including batch means, spectral variance estimators, and the initial convex sequence estimator.

Recommendations

• Estimation of the Asymptotic Variance in the CLT for Markov Chains
• Weighted batch means estimators in Markov chain Monte Carlo
• Strong consistency of multivariate spectral variance estimators in Markov chain Monte Carlo
• Variance estimation in the central limit theorem for Markov chains
• Kernel estimators of asymptotic variance for adaptive Markov chain Monte Carlo

Cites work

• scientific article; zbMATH DE number 3122097 (Why is no real title available?)
• scientific article; zbMATH DE number 3854249 (Why is no real title available?)
• scientific article; zbMATH DE number 193111 (Why is no real title available?)
• scientific article; zbMATH DE number 3561857 (Why is no real title available?)
• scientific article; zbMATH DE number 1324223 (Why is no real title available?)
• scientific article; zbMATH DE number 1022519 (Why is no real title available?)
• scientific article; zbMATH DE number 2117879 (Why is no real title available?)
• scientific article; zbMATH DE number 3357844 (Why is no real title available?)
• scientific article; zbMATH DE number 3390139 (Why is no real title available?)
• Adaptive bandwidth choice
• Autocorrelation, autoregression and autoregressive approximation
• Batch means and spectral variance estimators in Markov chain Monte Carlo
• Batch size selection for variance estimators in MCMC
• Bayesian Analysis of Binary and Polychotomous Response Data
• Completely monotone functions and sequences
• Convergence properties of Gibbs samplers for Bayesian probit regression with proper priors
• Covariance structure of the Gibbs sampler with applications to the comparisons of estimators and augmentation schemes
• Equation of state calculations by fast computing machines
• Estimating accuracy of the MCMC variance estimator: asymptotic normality for batch means estimators
• Estimation of a \(k\)-monotone density: limit distribution theory and the spline connection
• Estimation of a discrete monotone distribution
• Estimation of a discrete probability under constraint of \(k\)-monotonicity
• Fixed-Width Output Analysis for Markov Chain Monte Carlo
• Geometric convergence and central limit theorems for multidimensional Hastings and Metropolis algorithms
• Geometric ergodicity and hybrid Markov chains
• Geometric ergodicity and the spectral gap of non-reversible Markov chains
• Geometric ergodicity of Metropolis algorithms
• Handbook of Markov Chain Monte Carlo
• Least squares estimation of a completely monotone pmf: from analysis to statistics
• Markov Chains and Stochastic Stability
• Markov chain Monte Carlo: can we trust the third significant figure?
• Markov-chain monte carlo: Some practical implications of theoretical results
• Marshall lemma in discrete convex estimation
• Mixtures of exponential distributions
• Moment Problems for Compact Sets
• Monte Carlo error estimation for multivariate Markov chains
• Monte Carlo sampling methods using Markov chains and their applications
• Multivariate initial sequence estimators in Markov chain Monte Carlo
• Multivariate output analysis for Markov chain Monte Carlo
• Nonparametric estimation of species richness using discrete \(k\)-monotone distributions
• Nonparametric species richness estimation under convexity constraint
• Note on Completely Monotone Densities
• On multiply monotone distributions, continuous or discrete, with applications
• On the Markov chain central limit theorem
• On the geometric ergodicity of Hamiltonian Monte Carlo
• On variance conditions for Markov chain CLTs
• Relative fixed-width stopping rules for Markov chain Monte Carlo simulations
• Semiparametric Efficiency in Convexity Constrained Single-Index Model
• Strong Consistency and Other Properties of the Spectral Variance Estimator
• Strong consistency of multivariate spectral variance estimators in Markov chain Monte Carlo
• The Support Reduction Algorithm for Computing Non‐Parametric Function Estimates in Mixture Models
• The asymptotic validity of sequential stopping rules for stochastic simulations
• The geometry of mixture likelihoods: A general theory
• Time series: Theory and methods
• Uniform convergence of autocovariances
• Variable transformation to obtain geometric ergodicity in the random-walk Metropolis algorithm

Cited in

(2)

• Convergence rates of Metropolis-Hastings algorithms
• Multivariate strong invariance principles in Markov chain Monte Carlo

This page was built for publication: Efficient shape-constrained inference for the autocovariance sequence from a reversible Markov chain

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6183871)