基于P2P的混合过滤文章推荐系统

A Document Recommendation System Based on Clustering P2P Networks

研究了在分布式的情况下的混合推荐机制,根据P2P环境下文档线性增长的特点,提出了用户兴趣模板的更新算法。实验表明,基于混合过滤的方法推荐效果要远远好于基于内容过滤或基于协作过滤的方法。

This paper presents a document recommendation system based on clustering peer-to-peer networks. It’s an unstructured P2P system. In this system each agent-peer can learn user’s interest, then it helps user share and recommend documents with the other users. Since each peer in our P2P networks is a node, in order to cluster them, we import the concept of Group. Each group is composed of peers. The types of documents, which belong to a same group, are uniform. This paper presents how these peers help users to share and to recommend documents, and how they cluster into groups. Our experiment results show the advantages of the document recommendation system.