关键词: 协同过滤, 个性化推荐, 同行评议

Abstract: Employing the idea of collaborative filtering in personalized recommendation, this paper proposes a recommendation-propagation system of scientific achievements in which users’similarities are computed based on peer review. Through this recommendation-propagation system, each researcher can recommend his/her or other’s scientific literatures to the nearest neighbors (those researchers with similar research interests) in scientific community networks. By this way, effectively obtaining and actively introducing scientific achievements are interwoven by the peer collaborative filtering recommendation mechanism. Based on multi-agent simulation, this paper validates the performance of the recommendation- propagation system.