Arguing the Validation of Dunn's Index in Gene Clustering

Abstract

Cluster validation techniques are essential tools within cluster analysis, helpful to the interpretation of clustering results. In this study, the validation ability of Dunn's index in gene clustering was investigated with public gene expression datasets clustered by hierarchical clustering, K-means and Self-organizing maps. It was made clear that Dunn's index would give misleading validity results for its high sensitivity to noises in the data and should not be directly used in gene clustering validation.