Title: A selective survey of mixture models with flexible distributions
Authors: Sharon Lee - University of Queensland (Australia) [presenting]
Abstract: Extensive growth in the literature on multivariate flexible distributions in the past few decades has provided an abundance of choices for modelling the distribution of non-normal data. Some of these have been applied to mixture models and cluster analyses, in particular those that can accommodate skewness and heavy-tailedness. We provide an updated survey on recent developments in this area, focusing on the more popular proposals and those with publicly available software implementations. The different formulations used to construct these distributions may render them suitable for different applications. Their performances at modelling different types of asymmetric data are studied using simulations.