Abstract: The search content of node Peer-to-Peer(P2P) users reflects the user interest. So this paper proposes a search mechanism based on keyword relationship and node interest. The mechanism learns focus on the relationship between keywords, finds node by user interest. The use of keywords in the search increases the target semantic relations, and improves the search success rate. According to the node of interest, it narrows search and reduces the search overhead. Experimental results show that the fast index table method and the feedback mechanism can effectively improve the search performance.

Key words: Peer-to-Peer(P2P) network, node interest, keyword relationship, information retrieval

摘要： P2P节点用户的搜索内容在一定程度上体现了用户的兴趣。为此，提出一种基于关键字关联和节点兴趣的P2P信息搜索机制。该机制注重关键字之间关系的学习，通过用户操作发现节点兴趣。在搜索中利用关键字的语义关系增加命中目标，提高搜索成功率。根据节点兴趣缩小搜索范围、降低搜索开销。实验结果表明，采用快速建立索引表的方法和反馈机制能有效提高搜索性能。

关键词: P2P网络, 节点兴趣, 关键字关联, 信息搜索

CLC Number: 

• TP311