Some issues in robust clustering

TL;DR

This paper discusses key challenges in robust clustering, especially with Gaussian mixture models, including outlier definition, cluster-outlier ambiguity, and the impact of tuning decisions on clustering stability and estimation of cluster numbers.

Contribution

It highlights critical issues in robust clustering methods, emphasizing the need for clearer outlier definitions and better stability measurements, advancing understanding of Gaussian mixture model-based clustering.

Findings

Identifies ambiguity between outliers and clusters

Highlights dependence on tuning parameters

Critiques current stability measurement methods

Abstract

Some key issues in robust clustering are discussed with focus on Gaussian mixture model based clustering, namely the formal definition of outliers, ambiguity between groups of outliers and clusters, the interaction between robust clustering and the estimation of the number of clusters, the essential dependence of (not only) robust clustering on tuning decisions, and shortcomings of existing measurements of cluster stability when it comes to outliers.