Concentration graph models are an interesting framework to explore the conditional independence structure in a multivariate Normal distribution. The SINFUL procedure (Drton, 2004) is an appealing method for model selection in these kinds of models. In this paper, we propose a simple way to robustify SINFUL procedure to the presence of outliers. The solution proposed is based on the minimum covariance determinant estimator for the variance-covariance matrix. A simulation study confirmed that our choice is a useful way to robustify the estimates and the model selection procedure in concentration graph models.

Robustifying SINful procedure / Gottard, Anna; Pacillo, S.. - STAMPA. - (2007), pp. 577-580. (Intervento presentato al convegno CLADAG 2007 tenutosi a Università di Macerata).

Robustifying SINful procedure

GOTTARD, ANNA;
2007

Abstract

Concentration graph models are an interesting framework to explore the conditional independence structure in a multivariate Normal distribution. The SINFUL procedure (Drton, 2004) is an appealing method for model selection in these kinds of models. In this paper, we propose a simple way to robustify SINFUL procedure to the presence of outliers. The solution proposed is based on the minimum covariance determinant estimator for the variance-covariance matrix. A simulation study confirmed that our choice is a useful way to robustify the estimates and the model selection procedure in concentration graph models.
2007
Book of Short Papers
CLADAG 2007
Università di Macerata
Gottard, Anna; Pacillo, S.
File in questo prodotto:
File Dimensione Formato  
cladag07.pdf

Accesso chiuso

Tipologia: Pdf editoriale (Version of record)
Licenza: Tutti i diritti riservati
Dimensione 184.92 kB
Formato Adobe PDF
184.92 kB Adobe PDF   Richiedi una copia

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

Utilizza questo identificatore per citare o creare un link a questa risorsa: https://hdl.handle.net/2158/326534
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact