An overview of skew distributions in model-based clustering

The literature on non-normal model-based clustering has continued to grow in recent years. The non-normal models often take the form of a mixture of component densities that offer a high degree of flexibility in distributional shapes. They handle skewness in different ways, most typically by introducing latent 'skewing' variable(s), while some other consider marginal transformations of the original variable(s). We provide a selective overview of the main types of skew distributions used in the area, based on their characterization of skewness, and discuss different skew shapes they can produce. For brevity, we focus on the more commonly-used families of distributions. Crown Copyright (C) 2021 Published by Elsevier Inc. All rights reserved.

Automated summary

The literature on non-normal model-based clustering has been expanding in recent years. These models often use a mixture of component densities to provide flexibility in distributional shapes and handle skewness. Skewing is typically achieved by introducing latent variables or considering marginal transformations of the original variables.