Variational Bayes Approximations for Clustering via Mixtures of Normal Inverse Gaussian Distributions

TL;DR

This paper introduces a variational Bayes method for parameter estimation in clustering models based on mixtures of normal inverse Gaussian distributions, offering computational advantages over traditional EM algorithms.

Contribution

It presents a novel variational Bayes approach for NIG mixture models, reducing computational complexity and uncertainty compared to EM-based methods.

Findings

Effective on simulated data

Successful application to real data

Reduces computational time

Abstract

Parameter estimation for model-based clustering using a finite mixture of normal inverse Gaussian (NIG) distributions is achieved through variational Bayes approximations. Univariate NIG mixtures and multivariate NIG mixtures are considered. The use of variational Bayes approximations here is a substantial departure from the traditional EM approach and alleviates some of the associated computational complexities and uncertainties. Our variational algorithm is applied to simulated and real data. The paper concludes with discussion and suggestions for future work.