PS-Tree:一种基于路径的语义文档索引
PS-Tree:a Path-based Semantic Document Indexing
-
摘要: 针对以多个关键词形成的节点集合作为输入的语义路径包含查询问题,提出了PS-Tree结构,并将其应用到语义搜索引擎的索引结构上.PS-Tree借鉴了面向对象数据库中的集合值属性索引的方法,通过范围查询快速地遍历树形索引,将用户的多关键词查询输入请求对应到包含其语义连接的语义路径上,实现以节点集合作为输入的查询功能.实验结果表明,当查询关键词之间的语义距离较长时,PS-Tree结构表现出良好的性能.Abstract: In order to resolve the issue of containing query with a node set formed by multi-keywords,this paper proposes an index structure,named PS-Tree,and applies it to a semantic search engine.The PS-Tree refers to a set attribute index of object-orient database.It can map user's query to a semantic path which contains semantic relations as the aim of user's query in form of multi-Key words by "range query".It can improve the performance of index structure.The experiment result shows that the PS-Tree gives a good performance especially when there is a long semantic distance between the query Key words.