Title: Summarizing posterior clustering distributions
Authors: David Dahl - Brigham Young University (United States) [presenting]
Devin Johnson - Brigham Young University (United States)
Peter Mueller - UT Austin (United States)
Abstract: The aim is to address the problem of point estimation of a clustering distribution based on posterior samples, as well as the assessment of clustering uncertainty. We both extend the literature of loss functions for Bayesian clustering and also introduce a fast, scalable optimization procedure to obtain an optimal Bayesian estimate. Our approach is a stochastic search based on a series of micro-optimizations performed in random order and is embarrassingly parallel. We explain the algorithm and computational shortcuts, demonstrate the software, and compare its performance against increasingly-challenging applications of Bayesian clustering.