Discriminative Ordering Through Ensemble Consensus

TL;DR

This paper introduces a novel ensemble clustering method that constructs a discriminative ordering based on the distance between individual models and a consensus matrix, effectively ranking models that best match the consensus.

Contribution

It proposes a new scoring approach for evaluating diverse clustering models within ensemble clustering, accommodating constraints and varying cluster counts.

Findings

The proposed score accurately ranks models matching the consensus in synthetic tests.

It outperforms existing scoring methods in comparing diverse clustering algorithms.

The method is compatible with clustering constraints and varying numbers of clusters.

Abstract

Evaluating the performance of clustering models is a challenging task where the outcome depends on the definition of what constitutes a cluster. Due to this design, current existing metrics rarely handle multiple clustering models with diverse cluster definitions, nor do they comply with the integration of constraints when available. In this work, we take inspiration from consensus clustering and assume that a set of clustering models is able to uncover hidden structures in the data. We propose to construct a discriminative ordering through ensemble clustering based on the distance between the connectivity of a clustering model and the consensus matrix. We first validate the proposed method with synthetic scenarios, highlighting that the proposed score ranks the models that best match the consensus first. We then show that this simple ranking score significantly outperforms other scoring methods when comparing sets of different clustering algorithms that are not restricted to a fixed number of clusters and is compatible with clustering constraints.