Abstract

This paper describes and illustrates new features of the NIMBLE computing environment (de Valpine et al., 2017) that enable simulation-based inference for general nonparametric Bayesian mixture models.