How many clusters? An information theoretic perspective

TL;DR

This paper introduces an information theoretic approach to determine the optimal number of clusters in data by considering a temperature parameter that balances detail and noise, avoiding external goodness criteria.

Contribution

It proposes a novel method that uses a statistical mechanics perspective to identify the maximum meaningful number of clusters based on data size and sampling bias.

Findings

Finite data sets impose limits on resolvable structure.

Optimal clustering temperature depends on data size.

Method effectively finds the maximum number of meaningful clusters.

Abstract

Clustering provides a common means of identifying structure in complex data, and there is renewed interest in clustering as a tool for the analysis of large data sets in many fields. A natural question is how many clusters are appropriate for the description of a given system. Traditional approaches to this problem are based either on a framework in which clusters of a particular shape are assumed as a model of the system or on a two-step procedure in which a clustering criterion determines the optimal assignments for a given number of clusters and a separate criterion measures the goodness of the classification to determine the number of clusters. In a statistical mechanics approach, clustering can be seen as a trade--off between energy-- and entropy--like terms, with lower temperature driving the proliferation of clusters to provide a more detailed description of the data. For finite data sets, we expect that there is a limit to the meaningful structure that can be resolved and therefore a minimum temperature beyond which we will capture sampling noise. This suggests that correcting the clustering criterion for the bias which arises due to sampling errors will allow us to find a clustering solution at a temperature which is optimal in the sense that we capture maximal meaningful structure -- without having to define an external criterion for the goodness or stability of the clustering. We show that, in a general information theoretic framework, the finite size of a data set determines an optimal temperature, and we introduce a method for finding the maximal number of clusters which can be resolved from the data in the hard clustering limit.