Keyword Extraction Based Peer Clustering

Peer clustering plays an important role in P2P systems like peer discovery, resource sharing and management, etc. Keywords provide rich semantic information about the peers' interests. Keyword extraction from documents is a useful method in topic retrieval and document clustering. Peers exchange resources and some of them are text documents like news and novels. Such documents represent the interests of a peer. This paper proposes a method for clustering peers using the exchange text documents between them. The documents are used for keyword extraction. The keyword extraction is treated as a decision problem and based on Bayesian decision theory. The peers' similarities can be calculated by keyword similarities. And the cluster method is based on the peers' similarities. The experiment gives satisfied results. Finally, the conclusion is discussed.