大规模本体分块与映射研究  

Research on Large Scale Ontology Partition and Mapping

作　　者：徐德智[1] 赖雅[1] 王润梅[1] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083

出　　处：《小型微型计算机系统》2012年第3期542-547,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60970096)资助

摘　　要：针对当前在大规模本体映射方面存在的不足,提出一种新的基于遗传算法的大规模本体分块与映射方法.首先,对本体进行预处理,把本体表示成有向无环图,将本体分块问题转换成图分割问题.然后采用基于遗传算法的GPO(Genetic-Parti-tion-Ontology)算法对有向无环图进行分块,该算法不需要输入分块个数;最后利用基于本体块结构和基于参考点相结合方法找到正确的块映射.理论分析和实验结果表明,本文提出的映射方法能取得很好的结果.Due to the shortcomings in current large scale ontology mapping,proposesd a new method for large ontology partition and mapping based on genetic algorithm.Firstly,the ontologys are represented in DAG（Directed Acyclic Graph）structures by preprocessing.So,the ontology partition question is turned into a graph partition problem.Then we use GPO（Genetic-Partition-Ontology） algorithm which based on genetic algorithm to partition the DGA,and the algorithm don′t need the parameter of block number.Finaly,blocks from different ontology are matched by combining two methods of ontology blocks structure as well as predifind anchors.Theoretical analysis and experimental results shows that the method proposed by this paper can acquire good matching results.

关 键 词：大规模本体 遗传算法 本体分块 块映射 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]