The Forward Search (FS) represents a useful tool for clustering data that include outlying observations, because it provides a robust clustering method in conjunction with an outlier identification method. In this paper we propose to rephrase the FS clustering process in the framework of mixture models. Through this alternative formulation, some recent theoretical results about FS-based outlier detection in multivariate normal data can be exploited and extended to the multiple cluster setting.
Some developments in forward search clustering / Calò D. G.. - STAMPA. - (2008), pp. 65-68. (Intervento presentato al convegno First Joint Meeting of the Société Francophone de Classification and the Classification and Data Analysis Group of the Italian Statistical Society tenutosi a Caserta, Italy nel 11-13 giugno 2008).
Some developments in forward search clustering
CALO', DANIELA GIOVANNA
2008
Abstract
The Forward Search (FS) represents a useful tool for clustering data that include outlying observations, because it provides a robust clustering method in conjunction with an outlier identification method. In this paper we propose to rephrase the FS clustering process in the framework of mixture models. Through this alternative formulation, some recent theoretical results about FS-based outlier detection in multivariate normal data can be exploited and extended to the multiple cluster setting.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.