On Monte Carlo methods for estimating ratios of normalizing constants

摘要: Recently, estimating ratios of normalizing constants has played an important role in Bayesian computations. Applications arise many aspects statistical inference. In this article, we present overview and discuss the current Monte Carlo methods for constants. Then propose a new ratio importance sampling method establish its theoretical framework. We find that can be better than methods, example, bridge (Meng Wong) path (Gelman Meng), sense minimizing asymptotic relative mean-square errors estimators. An example is given illustrative purposes. Finally, two special applications general implementation issues