Peer-based relay scheme of collaborative filtering for research literature

Abstract

Much work has been done in both industry and academia on filtering for research literature, however most existing studies have their limitations in coping with the inherent characteristics of research literature search, i.e., most articles attract very few readers among all the researchers, and the recommendations are often circulated through members of particular communities. In this paper we propose a peer-based relay scheme of collaborative filtering for, but not limited to research literature. In the scheme, a recommendation request is relayed through a social structure dynamically formed by co-peers with common interests, and the recommendation results are adjusted and propagated by the co-peers. A hybrid filtering approach is deployed in the scheme. © 2011 Springer-Verlag.