Agglomerative and divisive hierarchical Bayesian clustering

• Published in: Computational statistics & data analysis Vol. 176; p. 107566
• Main Authors: Burghardt, Elliot, Sewell, Daniel, Cavanaugh, Joseph
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01-12-2022
• Subjects: Dirichlet concentration parameter; Dirichlet distribution; Finite mixture model; Hierarchical clustering algorithms; Dirichlet concentration parameter; Finite mixture model; Hierarchical clustering algorithms; Dirichlet distribution
• ISSN: 0167-9473; 1872-7352
• DOI: 10.1016/j.csda.2022.107566

Cluster analysis methods are designed to discover groups of subjects or objects in datasets by uncovering latent patterns in data. Two model-based Bayesian hierarchical clustering algorithms are presented—divisive and agglomerative—that return nested clustering configurations and provide guidance on the plausible number of clusters in a principled way. These algorithms outperform many existing clustering methods on benchmark data. The methods are applied to identify subpopulations among Parkinson's disease subjects using only baseline data, and differing patterns of progression in the few years following diagnosis are demonstrated in the identified clusters.
•This paper presents two novel Bayesian model-based hierarchical clustering methods.•These approaches are flexible, not limited by data type or the conjugacy of priors.•Methods provide a metric to guide users to reasonable locations to cut dendrograms.•Methods are most useful with nested structure of clusters and mixed types of data.