Abstract
This paper analyzes the performance of importance sampling distributions for computing expectations with respect to a whole family of probability laws in the context of Markov chain Monte Carlo simulation methods. Motivations for such a study arise in statistics as well as in statistical physics. Two choices of importance sampling distributions are considered in detail: mixtures of the distributions of interest and distributions that are "uniform over energy levels" (motivated by physical applications). We analyze two examples, a "witch's hat" distribution and the mean field Ising model, to illustrate the advantages that such simulation procedures are expected to offer in a greater generality. The connection with the recently proposed simulated tempering method is also examined.
Citation
Neal Madras. Mauro Piccioni. "Importance sampling for families of distributions." Ann. Appl. Probab. 9 (4) 1202 - 1225, November 1999. https://doi.org/10.1214/aoap/1029962870
Information