Diagnostic tools of approximate Bayesian computation using the coverage property

TL;DR

This paper introduces diagnostic tools based on the coverage property to assess and improve the choice of the tolerance parameter  in approximate Bayesian computation, ensuring credible intervals have correct coverage.

Contribution

It develops theoretical and practical diagnostics for  selection in ABC, enhancing the reliability of posterior approximations.

Findings

Coverage diagnostics can identify appropriate  values.

The methodology improves the accuracy of ABC posterior intervals.

Application to human demographic history demonstrates practical utility.

Abstract

Approximate Bayesian computation (ABC) is an approach for sampling from an approximate posterior distribution in the presence of a computationally intractable likelihood function. A common implementation is based on simulating model, parameter and dataset triples, (m,\theta,y), from the prior, and then accepting as samples from the approximate posterior, those pairs (m,\theta) for which y, or a summary of y, is "close" to the observed data. Closeness is typically determined though a distance measure and a kernel scale parameter, \epsilon. Appropriate choice of \epsilon is important to producing a good quality approximation. This paper proposes diagnostic tools for the choice of \epsilon based on assessing the coverage property, which asserts that credible intervals have the correct coverage levels. We provide theoretical results on coverage for both model and parameter inference, and adapt these into diagnostics for the ABC context. We re-analyse a study on human demographic history to determine whether the adopted posterior approximation was appropriate. R code implementing the proposed methodology is freely available in the package "abc."