Discussion on Bayesian Cluster Analysis: Point Estimation and Credible Balls by Sara Wade and Zoubin Ghahramani

TL;DR

This paper discusses Bayesian cluster analysis, focusing on point estimation, credible balls, interpretability as confidence balls, computational efficiency, and theoretical properties like posterior consistency.

Contribution

It provides insights into the interpretation of credible balls as confidence sets and discusses methods to improve computational efficiency in Bayesian clustering.

Findings

Credible balls can be interpreted as confidence sets under certain conditions.

Proposed methods to reduce computational costs in Bayesian clustering.

Analysis of posterior consistency and contraction rates in the context of Bayesian cluster analysis.

Abstract

I begin my discussion by giving an overview of the main results. Then I proceed to touch upon issues about whether the credible ball constructed can be interpreted as a confidence ball, suggestions on reducing computational costs, and posterior consistency or contraction rates.