Multivariate Output Analysis for Markov chain Monte Carlo

TL;DR

This paper introduces a multivariate framework for determining when to stop Markov chain Monte Carlo simulations by estimating the multivariate effective sample size and providing bounds on the required sample size for desired precision.

Contribution

It develops a multivariate effective sample size estimator and links it to stopping rules, addressing the largely ignored multivariate aspect of Monte Carlo error in MCMC.

Findings

Proposes a strongly consistent multivariate covariance estimator.

Derives a lower bound on effective sample size for precision.

Demonstrates finite sample performance in various examples.

Abstract

Markov chain Monte Carlo (MCMC) produces a correlated sample for estimating expectations with respect to a target distribution. A fundamental question is when should sampling stop so that we have good estimates of the desired quantities? The key to answering this question lies in assessing the Monte Carlo error through a multivariate Markov chain central limit theorem (CLT). The multivariate nature of this Monte Carlo error largely has been ignored in the MCMC literature. We present a multivariate framework for terminating simulation in MCMC. We define a multivariate effective sample size, estimating which requires strongly consistent estimators of the covariance matrix in the Markov chain CLT; a property we show for the multivariate batch means estimator. We then provide a lower bound on the number of minimum effective samples required for a desired level of precision. This lower bound depends on the problem only in the dimension of the expectation being estimated, and not on the underlying stochastic process. This result is obtained by drawing a connection between terminating simulation via effective sample size and terminating simulation using a relative standard deviation fixed-volume sequential stopping rule; which we demonstrate is an asymptotically valid procedure. The finite sample properties of the proposed method are demonstrated in a variety of examples.