基于二部图模型的大本体分块与映射方法  

Large-scale Ontology Partition and Mapping Method Based on Bipartite Graph Model

作　　者：郭维[1] 徐德智[1] 邹亮[1] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083

出　　处：《计算机工程》2012年第23期37-41,46,共6页Computer Engineering

基　　金：国家自然科学基金资助项目(60970096)

摘　　要：针对当前大本体分块与映射方法中的块边界信息丢失严重、块间匹配计算复杂等问题,提出一种基于二部图模型的大本体分块与映射方法。根据本体概念相似度在2个本体间建立带权二部图模型,对该二部图进行分割,实现本体的关联分块,在子图中求解得到二部图的最佳匹配,并通过映射修正获得块内实体映射。实验结果表明,该方法的分块质量与映射效率均较优。Aiming at the defects of partition quality and mapping efficiency existed in large scale ontology mapping,this paper proposes a new method for large scale ontology partition and mapping based on bigraph graph model.A weighted bigraph is built by the similarity between the concepts of two ontologies and the bigraph is partitioned into subgraphs to achieve ontology co partition.Then the ontology mapping problem is translated into an optimal matching problem in sub graphs.It applies the revised rules to obtain alignments in the optimal matching.Experimental results show that partition quality and mapping efficiency of this method is well.

关 键 词：大本体 二部图 分块 映射 最佳匹配 映射修正 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]