KK Berthlesen, LA Breyer and GO Roberts
Perfect Posterior Simulation for Mature and Hidden Markov Models
Abstract: In this paper we present an application of read-once coupling from the past to problems in Bayesian inference for latent statistical models. We describe a method of simulating perfectly from the posterior distribution of the unknown mixture weights in a mixture model. Our method is extended to a more general mixture problem where unknown parameters exist for the mixture components, and to a hidden Markov model.