Enhancing a navigational based recommender system by link prediction methods - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Rapport Année : 2009

Enhancing a navigational based recommender system by link prediction methods

Résumé

Recommender systems are widely used for personalization of information on the web and information retrieval systems. Collaborative Filtering (CF) is the most popular recommen- dation technique. However, Classical CF (CCF) systems use only direct links and common features to model relationships between users. This report presents a Behavioral Network Collaborative Filtering approach (BNCF) that uses navi- gational patterns to model relationships between users and exploits social networks techniques, such as prediction link methods, to explore additional links throughout the behav- ioral network. The final aim consists in involving these new links in prediction generation, to improve recommendations quality. The proposed approach is evaluated in terms of ac- curacy on a real usage dataset. The experimentation shows the benefit of exploiting new links to compute predictions in terms of HMAE. Besides, the evaluation of a combined model (that exploits both best accurate BNCF and CCF based models) shows the importance of combining similari- ties of two different networks and its impact on accuracy of high predictions.
Fichier non déposé

Dates et versions

inria-00395571 , version 1 (15-06-2009)

Identifiants

  • HAL Id : inria-00395571 , version 1

Citer

Ilham Esslimani, Armelle Brun, Anne Boyer. Enhancing a navigational based recommender system by link prediction methods. [Intern report] 2009. ⟨inria-00395571⟩
134 Consultations
0 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More