Spying on the prior of the number of data clusters and the partition distribution in Bayesian cluster analysis

Cluster analysis aims at partitioning data into groups or clusters. In applications, it is common to deal with problems where the number of clusters unknown. Bayesian mixture models employed in such applications usually specify a flexible prior that takes account uncertainty respect However, major empirical challenge involving use these characterisation induced on partitions. This work introduces an approach compute descriptive statistics partitions for three selected developed areas finite mixtures and nonparametrics. The proposed methodology involves computationally efficient enumeration in-sample (termed as ``data clusters'') determining first two moments symmetric additive characterising accompanying reference implementation made available R package 'fipp'. Finally, we illustrate through comparisons also discuss implications elicitation applications.