An objective approach to cluster validation

Bouguessa, Mohamed; Wang, Shengrui et Sun, Haojun (2006). « An objective approach to cluster validation ». Pattern Recognition Letters, 27(13), pp. 1419-1430.

Fichier(s) associé(s) à ce document :

Prévisualisation	PDF Télécharger (368kB)

Résumé

Cluster validation is a major issue in cluster analysis. Many existing validity indices do not perform well when clusters overlap or there is signi¯cant variation in their covariance structure. The contribution of this paper is twofold. First, we propose a new validity index for fuzzy clustering. Second, we present a new approach for the objective evaluation of validity indices and clustering algorithms. Our validity index makes use of the covariance structure of clusters, while the evaluation approach utilizes a new concept of overlap rate that gives a formal measure of the di±culty of distinguishing between overlapping clusters. We have carried out experimental studies using data sets containing clusters of di®erent shapes and densities and various overlap rates, in order to show how validity indices behave when clusters become less and less separable. Finally, the e®ectiveness of the new validity index is also demonstrated on a number of real-life data sets.

Type:	Article de revue scientifique
Mots-clés ou Sujets:	Fuzzy clustering, Validity index, Overlapping clusters, Overlap rate, Truthed data set.
Unité d'appartenance:	Faculté des sciences > Département d'informatique
Déposé par:	Mohamed Bouguessa
Date de dépôt:	29 mars 2016 14:05
Dernière modification:	20 avr. 2016 20:03
Adresse URL :	http://archipel.uqam.ca/id/eprint/8038

Modifier les métadonnées (propriétaire du document)

Statistiques