结构化P2P网络中关键字搜索算法

Keyword search algorithm of structured peer-to-peer networks

为了解决分布式哈希表(DHT)不支持模糊匹配和查找延迟长等问题,提出使用改进的超立方体互连圈结构建立逻辑的关键字搜索层.对于给定的文档和描述该文档的关键字集合,使用反向文档索引(IDF)技术选择文档的重要关键字,并映射到环标号,将描述文档的关键字集映射到立方体标号,有效地提取了关键字对于文档的重要性信息,并通过构造生成树来实施搜索操作.实验结果表明,与使用超立方体结构作为关键字搜索层的算法相比,该算法在查询关键字较少时,具有更好的性能.

Distributed hash table （DHT） was used to locate resources in structured peer-to-peer network, but there was high delay in DHT network because DHT couldn＇t support fuzzy matching. A solution was proposed that a logic keyword search layer was built by modified cube-connected cycle structure. For a document and its keyword set, inversed document frequency （IDF） was used to choose the top keywords for the document, the top keywords for the document was mapped to a cyclic index, the whole keyword set for a document was mapped to cube index, which method extract the information that which document was more important to a document, search was implemented by a spanning binomial tree. Experiments demonstrated that the MCCC search scheme worked more efficiently when the number of query keywords was comparatively little.