基于本体贝叶斯网络模型的本体映射方法研究  

A Method Based on the Model of OBN for Ontology Mapping

作　　者：张凌宇[1] 陈淑鑫[1] 

机构地区：[1]齐齐哈尔大学计算中心,黑龙江齐齐哈尔161006

出　　处：《科技通报》2016年第1期133-137,157,共6页Bulletin of Science and Technology

基　　金：国家自然科学基金(61204127);黑龙江省自然科学基金资助项目(F030503;F201336)

摘　　要：不同本体之间的异构问题严重地影响了本体之间的知识共享和重用。针对本体之间异构性的问题,本文提出一种基于本体贝叶斯网络模型(OBN)的本体映射方法(OM-OBN:Ontology Mappingbased on OBN)。在映射的过程中,方法OM-OBN首先将待映射的本体模型转换成OBN模型,并为OBN模型的每个节点建立一个相似度表,然后根据相似度表中的数据建立本体之间的映射关系,然后再利用已有的映射关系去更新相似度表,最后采用迭代映射策略,反复地执行建立映射和更新相似度表的过程,直到OM-OBN无法建立新的映射关系为止。实验结果表明:方法OM-OBN可以实现本体模型到OBN模型的转换,并有效地利用转换后的模型来建立本体之间的映射关系。因此,方法OM-OBN可以解决异构本体之间知识共享和重用等问题。The problem of heterogeneity between different ontologies seriously affects the sharing and reusing of knowledge among ontologies. Aiming at the problem of heterogeneity, this paper proposes a method for Ontology Mapping based on the model of OBN (Ontology-Bayesian Network), called OM-OBN. During the process of mapping, OM-OBN firstly transforms the ontologies to be mapped into OBNs, and creates similarity tables for the nodes in OBNs. Then, the mappings are created, according to the data in similarity tables. Then, the mappings are used to update similarity tables. Finally, the iterative strategy of mapping is carried out, which is composed of creating mappings and updating similarity tables, until new mappings can not be created by OM-OBN. Experimental results indicate that the method OM-OBN is able to transform the model of ontology into the model of OBN, and create mappings between ontologies effectively by the transformed model. Therefore, OM-OBN can be used to solve the problem of sharing and reusing between heterogeneous ontologies.

关 键 词：本体 本体贝叶斯网络模型 本体映射 迭代映射策略 

分 类 号：TP317.2[自动化与计算机技术—计算机软件与理论]