A new methodology for robust clustering without specifying in advance the underlying number of Gaussian clusters is proposed. The procedure is based on iteratively trimming, assessing the goodness of fit, and reweighting. The forward version of our procedure is initialized with a high trimming level and K = 1 populations. The procedure is then iterated throughout a fixed sequence of decreasing trimming levels. New observations are added at each step and, whenever a goodness of fit rule is not satisfied, the number of components K is increased. A stopping rule prevents our procedure from using outlying observations. Additional use of a backward criterion is discussed.

A robust clustering procedure with unknown number of clusters

Francesco Dotto
;
2018-01-01

Abstract

A new methodology for robust clustering without specifying in advance the underlying number of Gaussian clusters is proposed. The procedure is based on iteratively trimming, assessing the goodness of fit, and reweighting. The forward version of our procedure is initialized with a high trimming level and K = 1 populations. The procedure is then iterated throughout a fixed sequence of decreasing trimming levels. New observations are added at each step and, whenever a goodness of fit rule is not satisfied, the number of components K is increased. A stopping rule prevents our procedure from using outlying observations. Additional use of a backward criterion is discussed.
2018
9788891910233
File in questo prodotto:
Non ci sono file associati a questo prodotto.

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/11386/4766090
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact