Cluster analysis: A modern statistical review

Abstract

Cluster analysis is a big, sprawling field. This review paper cannot hope to fully survey the territory. Instead, it focuses on hierarchical agglomerative clustering, k‐means clustering, mixture models, and then several related topics of which any cluster analysis practitioner should be aware. Even then, this review cannot do justice to the chosen topics. There is a lot of literature, and often it is somewhat ad hoc. That is generally the nature of cluster analysis—each application requires a bespoke analysis. Nonetheless, clustering has proven itself to be incredibly useful as an exploratory data analysis tool in biology, advertising, recommender systems, and genomics.