Comment on "Improved mutual information measure for clustering, classification, and community detection"

TL;DR

This paper critiques a proposed reduced mutual information measure for clustering evaluation, demonstrating through empirical analysis that it does not outperform normalized mutual information and highlighting the importance of Kappa.

Contribution

The paper provides an empirical critique of the reduced mutual information measure, showing its limitations compared to normalized mutual information and validating the use of Kappa.

Findings

Reduced mutual information cannot handle clustering evaluation challenges better than NMI.

Empirical evidence shows reduced mutual information introduces more difficulties.

Kappa is validated as a necessary measure in this context.

Abstract

A recent article proposed reduced mutual information for evaluation of clustering, classification and community detection. The motivation is that the standard normalized mutual information (NMI) may give counter-intuitive answers under certain conditions and particularly when the number of clusters differs between the two divisions under consideration. The motivation makes sense. However, the examples given in the article are not accurate, and this comment discusses why. In addition, this comment also empirically demonstrates that the reduced mutual information cannot handle the difficulties of NMI and even brings more. The necessity of Kappa is also empirically validated in this comment.