Exact Finite Mixture Representations for Species Sampling Processes

Published:Dec 30, 2025 18:56
1 min read
ArXiv

Analysis

This paper provides a computationally efficient way to represent species sampling processes, a class of random probability measures used in Bayesian inference. By showing that these processes can be expressed as finite mixtures, the authors enable the use of standard finite-mixture machinery for posterior computation, leading to simpler MCMC implementations and tractable expressions. This avoids the need for ad-hoc truncations and model-specific constructions, preserving the generality of the original infinite-dimensional priors while improving algorithm design and implementation.

Reference

Any proper species sampling process can be written, at the prior level, as a finite mixture with a latent truncation variable and reweighted atoms, while preserving its distributional features exactly.