一种用于在Peer-to-Peer系统中自动形成社区的元数据

Peer-to-Peer(P2P)系统为文件共享提供了一个较为理想的平台.然而,由于缺乏中央的控制机制,使得全局信息难以获得,这给在该环境下进行信息检索带来了很大的困难.基于社区的信息检索方式可以帮助用户在社区中找到更有价值的信息,同时降低因为查询而带来的各种资源消耗.文中提出了用带权值的关键词列表(WeightedTermList,WTL)来作为自动形成P2P社区的元数据.实验结果证实了用WTL来判断节点(Peer)的相似性是非常有效的.通过对实验数据的分析,得到两个重要的结论.

A Useful Metadata for Automatic Community Formation in Peer-to-Peer Systems

Peer-to-Peer (P2P) system provides an appealing platform for file sharing for its high autonomy, reliability and scalability. However, it is difficult to retrieve information in P2P systems in the absence of central mechanism and global information. The community-based information retrieval could help users find out more valuable information in the community and reduce the resource cost of searching. This paper suggested a weighted term list (WTL) as the metadata to automatically form P2P communities. The experiments prove that WTL is effective in finding out similar peers. Two important empirical conclusions were obtained through further investigation of the experimental results.