P2P-based PVR Recommendation using Friends, Taste Buddies and Superpeers

Abstract

In this paper we present a novel distributed recommendation method based on exchanging similar playlists among taste buddies, which takes into account the limited availability of peers, lack of trust in P2P networks, and dynamic identities of peers, etc. Our preliminary experiment shows that only exchanging a small portion of similar playlists from taste buddies could lead to an efficient way to calculate recommendations within a context of P2P networks.