基于两阶段查询重写的XML近似查询算法  

Two-Phase Query Rewriting Based Approximate XML Query Algorithm

作　　者：衡星辰[1] 覃征[1] 邵利平[1] 曹玉辉[1] 高洪江[1] 

机构地区：[1]西安交通大学电子与信息工程学院,陕西西安710049

出　　处：《电子学报》2007年第7期1271-1278,共8页Acta Electronica Sinica

基　　金：国家973重点基础研究发展规划(No.2004CB719401)

摘　　要：提出了基于两阶段查询重写的XML近似查询算法.该算法不仅能够返回精确查询结果,而且能够返回带有相似度分值的近似结果序列.首先,通过模式重写策略,将原始查询树改写为多种XML DTD(文档类型定义)下的重写查询树,从而解决了XML数据的多样性带来的查询语义缺失问题,接着,利用基本变异操作得到的变异查询树对XML数据树完成精确嵌入,可将XML近似查询的问题转变为多棵变异查询树的精确查询问题,并给出了基于XML数据统计的相似度计算模型和Top-K问题求解的优化算法.最后,在汽车外形智能化设计的实验中表明该算法优于SSO算法.A two-phase query rewriting based approximate XML （extensible markup language） query algorithm is proposed. The algorithm can not only return the exact answers,but also return the sequence of approximated answers with similarity degree. Firstly, through the strategy of scheme rewriting, an original query tree is rewritten into query trees with different XML DTDs （document type definition） so as to solve the problem of semantic loss due to the heterogeneous XML data. Secondly, the transformed query trees derived by the sequences of the basic transformation operations are used to perform exact embedings into XML data tree,so as to transform the problem of approximate XML query into the problem of exact XML query for transformed query trees. Thirdly, XML data distribution statistics based similarity degree computing model and optimization algorithm for Top-K problem are given. Finally, the experiments of intelligent design of automobile shape show our algorithm outperforms the SSO algorithm.

关 键 词：XML近似查询 基本变异操作 变异查询树 模式重写 异质XML文档 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]