【作者】 张天；

【导师】 徐德智；

【作者基本信息】 中南大学 ， 计算机软件与理论， 2008， 硕士

【摘要】 语义Web的目标是使得Web上的信息具有计算机可以处理的语义。语义原生XML数据库系统(简称SNAX)的目的在于构建一个提供研究语义Web的各种理论、技术和方法的平台。本体映射技术对于SNAX中的本体查询、本体集成等都是至关重要的一部分,本文将着重对此进行研究。首先,本文对语义Web的研究背景和SNAX系统的研发目的进行了简介,并描述了SNAX系统的体系结构和本体映射子系统SNAXMapping2的模块结构。其次,针对当前本体映射系统对语义信息的利用不充分,导致结果的查准率和查全率偏低,难以满足实际应用的需求的问题,提出并实现了一种基于推理的本体映射修正算法。然后实现了该算法,并将其融合进本体映射子系统SNAXMapping2中,并利用国际组织OAEI提供的标准测试数据集benchmarks对系统进行了测试,同时利用查准率和查全率作为衡量标准对映射结果进行了统计分析。从实验结果来看,此算法能提高映射结果的查准率。最后,针对当前本体映射评价标准并不完全基于语义,造成对本体映射的评价不够精确的问题,提出了一个计算本体映射纯语义查准率和查全率的框架,并证明了在该框架下实现的任意查准率和查全率函数都必然是纯语义的。然后提出了一组面向应用的函数,用于在这个框架下计算本体映射的纯语义查准率和查全率。更多还原

【Abstract】 The goal of Semantic Web is to make Web data machine-processable. Semantic NAtive XML database system (SNAX) aims to build a platform for research on various theories, techniques and methods of Semantic Web. Ontology mapping is a crucial part of SNAX, which serves for ontology query and integration. This thesis focuses on this issue.Firstly, the purpose of researching and developing SNAX is briefly introduced, as well as research background of Semantic Web. Then the architecture of SNAX system and the module structure of ontology mapping sub-system SNAXMapping2 are described.Secondly, an ontology mapping amendment algorithm based on reasoning is proposed and implemented, aiming to solve the problem that current ontology mapping systems can not provide results with precision and recall high enough for practical application, due to their insufficient utilization of semantic information in ontologies. The algorithm is implemented and then integrated into the ontology mapping sub-system SNAXMapping2, and "benchmarks", a standard test dataset for ontology mapping provided by OAEI, is used to test the system. The precision and recall of mapping results are statistically analyzed. The results show our algorithm can improve the precision of mapping result.Finally, a framework for computing purely semantic precision and recall of ontology mapping is proposed, aiming to solve the problem that current ontology mapping measures are not purely semantic, which makes evaluation of ontology mappings not precise. Then it is proved that any precision and recall functions instantiated under this framework must be purely semantic, and a group of application-oriented functions are proposed, to instantiate purely semantic precision and recall of ontology mapping under this framework.更多还原