A note on target distribution ambiguity of likelihood-free samplers

2010/05/28 by Scott A. Sisson, S. A. Sisson, Gareth W. Peters +9
Computer Science · Mathematics · #Advanced Statistical Methods and Models #Bayesian Methods and Mixture Models #Statistical Methods and Inference #stat.CO

paper · pdf · doi:10.48550/arxiv.1005.5201

arxiv created 2010/05/28 · arxiv updated 2010/05/31

Abstract

Methods for Bayesian simulation in the presence of computationally intractable likelihood functions are of growing interest. Termed likelihood-free samplers, standard simulation algorithms such as Markov chain Monte Carlo have been adapted for this setting. In this article, by presenting generalisations of existing algorithms, we demonstrate that likelihood-free samplers can be ambiguous over the form of the target distribution. We also consider the theoretical justification of these samplers. Distinguishing between the forms of the target distribution may have implications for the future development of likelihood-free samplers.

Related

• Measuring the accuracy of likelihood-free inference
• Classification and Bayesian Optimization for Likelihood-Free Inference
• A Note on Using Discretized Simulated Data to Estimate Implicit Likelihoods in Bayesian Analyses
• Bayesian Optimization for Likelihood-Free Inference of Simulator-Based Statistical Models
• Approximate Bayesian inference from noisy likelihoods with Gaussian process emulated MCMC