A Social Recommender System With Privacy Preserving

Résumé: P2P social networks are presented to give the users more control of their data and relations. This is due to the decentralized underlying architecture. Several architectures were proposed. They had proved the feasibility of implementing basic social networks features (eg. publish, share, comment …) on top of P2P architectures. Advanced features as social information retrieval and particularly the recommender systems are very hard to implement. This is due to the absence of a central server that has a complete view of the social graph, it constitutes a recent challenge. In this paper, we propose a novel recommender system for P2P social network. The principal objective is to help users to discover the most relevant content shared by their friends. Performance evaluations show that our proposed model is globally satisfactory compared to existing similar solutions.

Publié dans la revue: Séminaire Mathématique de Béjaia