面向异质可扩展标记语言文档的近似查询算法

Approximate Query Algorithm for Heterogeneous-Oriented Extensible Markup Language Documents

提出了一种面向异质可扩展标记语言（XML）文档的近似查询算法（TPQR）．利用模式重写策略，通过对XML模式之间语义相似度及节点映射关系的求解，将原始查询树改写为多种XML模式下的重写查询树，从而解决了XML数据的多样性带来的查询语义缺失问题．将经变异操作序列得到的变异查询树精确地嵌入在XML数据树之中，即可将XML近似查询问题转变为多棵变异查询树的精确查询问题．在汽车外形智能化设计的实验中比较了TPQR算法和静态有序选择算法，结果表明，在查准率、查全率上，前者比后者分别提高了5％～10％和10%～30%．

An approximate query algorithm for heterogeneous-oriented extensible markup lan guage （XML） documents-TPQR was proposed. Based on the strategy of scheme rewriting, an original query tree was rewritten into the query trees with multi-class XML scheme through calculating the semantic similarity and nodes mapping relationship among XML schemes, as so to solve the problem of lack of query semantic due to the diversity of heterogeneous XML data. The mutated query trees derived from the sequences of the basic mutation operations complete the exact embedding for XML data tree, so that the problem of approximate XML query is transformed into an exact XML query problem of multi-mutation trees query. The experiments of the intelli gent design of automobile shape show that compared with SSO algorithm, the precision ratio and recall ratio of TPQR algorithm is increased by 5%- 10% and 10%-30% respectively.