Generalization of c-means for identifying non-disjoint clusters with overlap regulation

Abstract : Clustering is an unsupervised learning method that enables to fit structures in unlabeled data sets. Detecting overlapping structures is a specific challenge involving its own theoretical issues but offering relevant solutions for many application domains. This paper presents generalizations of the c-means algorithm allowing the parametrization of the overlap sizes. Two regulation principles are introduced, that aim to control the overlap shapes and sizes as regard to the number and the dispersal of the cluster concerned. The experiments performed on real world datasets show the efficiency of the proposed principles and especially the ability of the second one to build reliable overlaps with an easy tuning and whatever the requirement on the number of clusters.
Type de document :
Article dans une revue
Pattern Recognition Letters, Elsevier, 2014, pp.92-98. 〈10.1016/j.patrec.2014.03.007〉
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-00978269
Contributeur : Guillaume Cleuziou <>
Soumis le : lundi 5 mai 2014 - 16:37:26
Dernière modification le : jeudi 7 février 2019 - 17:11:52
Document(s) archivé(s) le : mardi 5 août 2014 - 11:00:39

Fichier

hal-00978269.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Citation

Chiheb-Eddine Ben N'Cir, Guillaume Cleuziou, Nadia Essoussi. Generalization of c-means for identifying non-disjoint clusters with overlap regulation. Pattern Recognition Letters, Elsevier, 2014, pp.92-98. 〈10.1016/j.patrec.2014.03.007〉. 〈hal-00978269〉

Partager

Métriques

Consultations de la notice

469

Téléchargements de fichiers

167