Comparison of asymptotic variances of inhomogeneous Markov chains with application to Markov chain Monte Carlo methods - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Annals of Statistics Année : 2014

Comparison of asymptotic variances of inhomogeneous Markov chains with application to Markov chain Monte Carlo methods

Résumé

In this paper, we study the asymptotic variance of sample path averages for inhomogeneous Markov chains that evolve alternatingly according to two different π-reversible Markov transition kernels P and Q. More specifically, our main result allows us to compare directly the asymptotic variances of two inhomogeneous Markov chains associated with different kernels Pi and Qi, i ∈ {0, 1}, as soon as the kernels of each pair (P0, P1) and (Q0, Q1) can be ordered in the sense of lag-one autocovariance. As an important application, we use this result for comparing different data-augmentation-type Metropolis- Hastings algorithms. In particular, we compare some pseudo-marginal algorithms and propose a novel exact algorithm, referred to as the random refreshment algorithm, which is more efficient, in terms of asymptotic variance, than the Grouped Independence Metropolis- Hastings algorithm and has a computational complexity that does not exceed that of the Monte Carlo Within Metropolis algorithm

Dates et versions

hal-01262407 , version 1 (26-01-2016)

Identifiants

Citer

Florian Maire, Randal Douc, Jimmy Olsson. Comparison of asymptotic variances of inhomogeneous Markov chains with application to Markov chain Monte Carlo methods. Annals of Statistics, 2014, 42 (4), pp.1483 - 1510. ⟨10.1214/14-AOS1209⟩. ⟨hal-01262407⟩
36 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More