Query expansion using fuzzy association rules between terms

C. Latiri-Cherif S. Ben-Yahia Jean-Pierre Chevallet 1 A. Jaoua
1 MRIM - Modélisation et Recherche d’Information Multimédia [Grenoble]
LIG - Laboratoire d'Informatique de Grenoble, Inria - Institut National de Recherche en Informatique et en Automatique
Abstract : Information retrieval (IR) focuses on the process of determining and assessing the adequacy between a user-query and a collection of documents, yielding a subset of relevant documents. In this respect, query expansion aims to reduce an eventual query/document mismatch by expanding the query using correlated terms. In this paper, we present an approach based on the use of association rules to detect such correlations, in order to improve retrieval effectiveness by reducing such mismatch. By considering the term-document relation as a fuzzy binary relation (with a special emphasis on membership degrees semantic), we propose a fuzzy conceptual approach to extract fuzzy association rules. This approach is based on the closure of an extended fuzzy Galois connection, using different semantics of term membership degrees. An experimental study, on real textual collections, has confirmed our intuitive hypothesis that the synergy between association rules and query expansion is fruitful. Results of the study show a significant improvement in the performances of the information retrieval system, both in terms of recall and precision.
Type de document :
Communication dans un congrès
JIM'2003 conference Journées de l'Informatique Messine, 2003, Metz, France. 2003
Liste complète des métadonnées

https://hal.inria.fr/hal-00953930
Contributeur : Marie-Christine Fauvet <>
Soumis le : vendredi 28 février 2014 - 16:06:53
Dernière modification le : jeudi 11 janvier 2018 - 06:22:06

Identifiants

  • HAL Id : hal-00953930, version 1

Collections

Citation

C. Latiri-Cherif, S. Ben-Yahia, Jean-Pierre Chevallet, A. Jaoua. Query expansion using fuzzy association rules between terms. JIM'2003 conference Journées de l'Informatique Messine, 2003, Metz, France. 2003. 〈hal-00953930〉

Partager

Métriques

Consultations de la notice

181