Recent results in model-based clustering via the Cluster‐Weighted approach

TItolo: Recent results in model-based clustering via the Cluster‐Weighted approach 

Relatore: Salvatore Ingrassia, Department of Economics and Business, University of Catania

Coordinate: mercoledì 17 febbraio, ore 11:30, Aula Ovest, Osservatorio Astrofisico di Catania

Abstract: Finite mixture models provide a flexible approach to statistical modeling of a wide variety of random phenomena characterized by unobserved heterogeneity. Cluster-weighted models (CWMs) are a flexible family of mixture models for modeling the relationship between a response variable and a set of covariates when data comes from a heterogeneous population.
In this talk, we introduce a wide family of CWMs where the component conditional distributions are assumed to belong to different probability distributions and the covariates are allowed to be of mixed-type (that is containing both categorical and continuous variables). Under the assumption of Gaussian covariates, sufficient conditions for model identifiability are provided. Moreover, maximum likelihood parameter estimates are derived using the EM algorithm.