Online Program Home
My Program

Abstract Details

Activity Number: 176 - Bayesian Mixture Modeling, Clustering and Unsupervised Learning
Type: Contributed
Date/Time: Monday, July 29, 2019 : 10:30 AM to 12:20 PM
Sponsor: Section on Bayesian Statistical Science
Abstract #306427
Title: Mixtures of Multivariate Skew Normal Generalised Hyperbolic Factor Analyzer Models in a Bayesian Framework
Author(s): Darren Wraith* and Mohsen Maleki
Companies: Queensland University of Technology and Shiraz University, Iran
Keywords: Factor analysis; Bayesian statistics; Asymmetric distributions; Mixture models

A mixture of factor analyzer (MFA) model provides an efficient tool for the analysis of high dimensional data by reducing the number of free parameters through its factor-analytic representation of the covariance matrices. The model also provides an important tool to identify hidden or latent groups in the data. Recent approaches to extend the model to allow for skewed and/or heavy tailed data have been examined in a frequentist setting where there are some known computational limitations. In this paper we introduce a MFA model based on the unrestricted skew normal generalized hyperbolic (SUNGH) distribution in a Bayesian framework where there are some computational advantages. The SUNGH family is a broad class of distributions providing considerable flexibility to model skewness in different directions as well as allowing for heavy tailed data. The SUNGH family also has several desirable properties, including an analytically tractable density and ease of computation for simulation and estimation of parameters. For factor analysis models, the SUNGH family also allows for skewness and heavy tails for both the error component and factor scores.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program