Title: A fresh consideration for the mode within a mixture context
Authors: JT Ferreira - University of Pretoria (South Africa) [presenting]
Abstract: With an astounding increase and necessitated understanding of scale and scope in modern data, flexible modelling remains a sincere and meaningful point of discussion within data analysis. Plentiful work has been studied based on mean/variance/mean-variance type of mixtures; a specific compound approach is of interest with a focus on the mode. Particular positive support distributions are considered in their known reparameterised form where the mode is explicitly available, and acts as a point of departure for studying the construction and viability of a mixture-type approach on this mode. This characterisation allows for unique leverage in terms of modelling by using the mode as an intuitive parameter that may allow for multiple bumps in data, as well as potential intuitive initialisation choices in estimation. The results are illustrated via a real data application.