基于本体的多源异构XML数据近似查询方法

Approximate Query Approach Based on Ontology for Multi-Source and Heterogeneous XML Data

提出了一种基于本体的面向多源异构可扩展置标语言（XML）的近似查询方法．该方法将传统的基于查询模式树松弛的近似查询策略与基于本体的XML数据集成思想相融合，首先建立文档类型定义结构与全局模式的映射模型（PTO），再将PTO与近似查询领域的松弛操作、打分机制相结合，提出了一种新的XML近似查询算法——OAXQ．这样，用户在全局模式下的查询语言不再是Xpath查询表达式，而是对象查询语言的一个简单子集，松弛的产生不再依靠查询模式树的变换，而是通过一系列映射规则转化而成．汽车外型的智能设计实验表明，OAXQ算法的查询速度比单纯基于查询模式树松弛策略的静态有序选择算法平均提高了20倍，查询准确度提高了2％～24％．

An ontology-based approach to effectively process approximate query for multi-source and heterogeneous XML data is proposed, in which the traditional approximate XML query strategy using tree pattern relaxation is integrated with the idea of ontology-based XML data integration. Firstly, a mapping model（PTO） from document type definition（DTD） structure to global schema is constructed, then PTO is combined with relaxation and scoring mechanism and a novel algorithm--OAXQ is presented. Thus, user query language based on global schema becomes a simple subset of object query language instead of Xpath expressions, and relaxations are accomplished by a series of conversion of mapping rules instead of transform of query tree patterns. Experiments of intelligent design of automobile shape show that comparing with SSO algorithm the query speed and precision is increased by 20 times and 20 %- 24% respectively by OAXQ