Abstract
AbstractThe problem of optimally scaling the proposal distribution in a Markov chain Monte Carlo algorithm is critical to the quality of the generated samples. Much work has gone into obtaining such results for various Metropolis–Hastings (MH) algorithms. Recently, acceptance probabilities other than MH are being employed in problems with intractable target distributions. There are few resources available on tuning the Gaussian proposal distributions for this situation. We obtain optimal scaling results for a general class of acceptance functions, which includes Barker’s and lazy MH. In particular, optimal values for Barker’s algorithm are derived and found to be significantly different from that obtained for the MH algorithm. Our theoretical conclusions are supported by numerical simulations indicating that when the optimal proposal variance is unknown, tuning to the optimal acceptance probability remains an effective strategy.
Publisher
Cambridge University Press (CUP)
Subject
Applied Mathematics,Statistics and Probability
Reference41 articles.
1. Diffusion limit for the random walk Metropolis algorithm out of stationarity;Kuntz;Ann. Inst. H. Poincaré Prob. Statist.,2019
2. Optimal scaling of random walk Metropolis algorithms using Bayesian large-sample asymptotics;Schmon;Statist. Comput.,2022
3. Hybrid Monte Carlo;Duane;Phys. Lett. B,1987
4. [23] Morina, G. , Łatuszyński, K. , Nayar, P. and Wendland, A. (2021). From the Bernoulli factory to a dice enterprise via perfect sampling of Markov chains. To appear in Ann. Appl. Prob.
5. Barker’s algorithm for Bayesian inference with intractable likelihoods;Gonçalves;Brazilian J. Prob. Statist.,2017