Evidential relational clustering using medoids

Abstract : In real clustering applications, proximity data, in which only pairwise similarities or dissimilarities are known, is more general than object data, in which each pattern is described explicitly by a list of attributes. Medoid-based clustering algorithms, which assume the prototypes of classes are objects, are of great value for partitioning relational data sets. In this paper a new prototype-based clustering method, named Evidential C-Medoids (ECMdd), which is an extension of Fuzzy C-Medoids (FCMdd) on the theoretical framework of belief functions is proposed. In ECMdd, medoids are utilized as the prototypes to represent the detected classes, including specific classes and imprecise classes. Specific classes are for the data which are distinctly far from the prototypes of other classes, while imprecise classes accept the objects that may be close to the prototypes of more than one class. This soft decision mechanism could make the clustering results more cautious and reduce the misclassification rates. Experiments in synthetic and real data sets are used to illustrate the performance of ECMdd. The results show that ECMdd could capture well the uncertainty in the internal data structure. Moreover, it is more robust to the initializations compared with FCMdd.
Type de document :
Communication dans un congrès
The 18th International Conference on Information Fusion, Jul 2015, Washington, United States
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01176143
Contributeur : Kuang Zhou <>
Soumis le : mardi 14 juillet 2015 - 23:06:07
Dernière modification le : mercredi 2 août 2017 - 10:06:18
Document(s) archivé(s) le : mercredi 26 avril 2017 - 04:01:37

Fichiers

sECMdd_fusion2015_revise.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01176143, version 1
  • ARXIV : 1507.04091

Citation

Kuang Zhou, Arnaud Martin, Quan Pan, Zhun-Ga Liu. Evidential relational clustering using medoids. The 18th International Conference on Information Fusion, Jul 2015, Washington, United States. 〈hal-01176143〉

Partager

Métriques

Consultations de
la notice

244

Téléchargements du document

48