A family of parsimonious Gaussian cluster-weighted models is presented. This family concerns a multivariate extension to cluster-weighted modelling that can account for correlations between multivariate responses. Parsimony is attained by constraining parts of an eigen-decomposition imposed on the component covariance matrices. A sufficient condition for identifiability is provided and an expectation-maximization algorithm is presented for parameter estimation. Model performance is investigated on both synthetic and classical real data sets and compared with some popular approaches. Finally, accounting for linear dependencies in the presence of a linear regression structure is shown to offer better performance, vis-à-vis clustering, over existing methodologies.

Multivariate response and parsimony for Gaussian cluster-weighted models

PUNZO, ANTONIO;INGRASSIA, Salvatore;
2017-01-01

Abstract

A family of parsimonious Gaussian cluster-weighted models is presented. This family concerns a multivariate extension to cluster-weighted modelling that can account for correlations between multivariate responses. Parsimony is attained by constraining parts of an eigen-decomposition imposed on the component covariance matrices. A sufficient condition for identifiability is provided and an expectation-maximization algorithm is presented for parameter estimation. Model performance is investigated on both synthetic and classical real data sets and compared with some popular approaches. Finally, accounting for linear dependencies in the presence of a linear regression structure is shown to offer better performance, vis-à-vis clustering, over existing methodologies.
2017
Cluster-weighted models; EM algorithm; Multivariate response; Model-based clustering; Mixture models; Parsimonious models
File in questo prodotto:
File Dimensione Formato  
Dang, Punzo, McNicholas, Ingrassia & Browne (2017) - JC.pdf

solo gestori archivio

Descrizione: Articolo principale
Tipologia: Versione Editoriale (PDF)
Dimensione 508.22 kB
Formato Adobe PDF
508.22 kB Adobe PDF   Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11769/50741
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 46
  • ???jsp.display-item.citation.isi??? 41
social impact