Mining Closed Gradual Patterns

Abstract : Mining gradual rules of the form −"the more A, the more B"− is more and more grasping the interest of the data mining commu- nity. Several approaches have been recently proposed. Unfortunately, in all surveyed approaches, reducing the quantity of mined patterns (and, consequently, the quantity of extracted rules) was not the main concern. To palliate such a drawback, a possible solution consists in using results of Formal Concept Analysis to generate a lossless reduced size nucleus of gradual patterns. To do so, we introduce in this paper a novel closure op- erator acting on gradual itemsets. Results of the experiments carried out on synthetic datasets showed important profits in terms of compactness of the generated gradual patterns set.
Type de document :
Communication dans un congrès
ICAISC: International Conference on Artificial Intelligence and Soft Computing, 2010, Zakopane, Poland. pp.267-274, 2010
Liste complète des métadonnées

https://hal-lirmm.ccsd.cnrs.fr/lirmm-00504029
Contributeur : Sarra Ayouni <>
Soumis le : lundi 19 juillet 2010 - 16:10:35
Dernière modification le : jeudi 11 janvier 2018 - 06:26:17

Identifiants

  • HAL Id : lirmm-00504029, version 1

Collections

Citation

Sarra Ayouni, Anne Laurent, Sadok Benyahia, Pascal Poncelet. Mining Closed Gradual Patterns. ICAISC: International Conference on Artificial Intelligence and Soft Computing, 2010, Zakopane, Poland. pp.267-274, 2010. 〈lirmm-00504029〉

Partager

Métriques

Consultations de la notice

55