Dynamic mixture of finite mixtures of factor analysers
File(s)
Published version (advance publication)
Author(s)
Grushanina, Margarita
Frühwirth-Schnatter, Sylvia
Type
Journal Article
Abstract
Mixtures of factor analysers represent a popular tool for finding structure in data. While in many applications the number of clusters and latent factors within clusters is held constant, some recent models automatically infer cluster and/or factor dimensionalities. This is done by employing nonparametric priors and allowing the number of clusters and factors to potentially be infinite. Markov chain Monte Carlo (MCMC) estimation is performed via adaptive algorithms, where parameters associated with the redundant factors are discarded. The current work contributes to the literature by allowing automatic inference on the number of clusters and cluster-specific factors while keeping both dimensions finite. For automatic inference on the cluster structure we employ the dynamic mixture of finite mixtures model with a prior on the number of mixture components. Automatic inference on cluster-specific factors is performed by assigning an exchangeable shrinkage process prior, which can be interpreted as a generalized cumulative shrinkage process prior for the columns of the factor loading matrices. Extensive simulation studies and applications to benchmark as well as real data sets demonstrate that our model outperforms competing alternatives, in particular those based on the multiplicative gamma process prior, with respect to recovering the correct number of cluster-specific factors.
Date Issued
2025-08-18
Date Acceptance
2025-08-01
Citation
Bayesian Analysis, 2025
ISSN
1936-0975
Publisher
International Society for Bayesian Analysis
Journal / Book Title
Bayesian Analysis
Copyright Statement
Rights: Copyright © 2025 International Society for Bayesian Analysis
License URL
Publication Status
Published online
Date Publish Online
2025-08-18