DeHinter: A Social-oriented Peer-to-Peer Recommender System

DeHinter is a Peer-to-Peer (P2P) recommender system that exploits social filtering techniques in order to implement a fully decentralized resource sharing platform. The system provides to users a way to share, search and retrieve contents in a scalable, flexible and efficient way. The spontaneous relationships between users that show similar interests shape highly connected thematic clusters that can be exploited to provide personalized advices. DeHinter's goal is to reduce the impact of the information overload providing a decentralized, autonomous and efficient way to filter contents exploiting social-oriented phenomena. The P2P communication layer of DeHinter is based on the P2P Gnutella protocol, that is a fully distributed overlay network. Main features: - Starts from the evidence that affinities between Gnutella users form scale-free and small world patterns. - Implement a distributed recommender engine able to suggest contents in a transparent way. - Fully decentralized and autonomous. - Exploits a de facto “word of mouth” mechanism. - Independent from content metadata descriptions or a specific ontology. - Users can give both implicit and explicit feedbacks. - Simple, completely anonymous and protecting users privacy mechanism to gather usage data for evaluation purposes.