一种自适应的多策略本体映射方法  

An adaptive multi-strategy ontology matching method

作　　者：孙煜飞[1] 马良荔[1] 周润芝 

机构地区：[1]海军工程大学电子工程学院 [2]91311部队

出　　处：《海军工程大学学报》2016年第2期75-80,85,共7页Journal of Naval University of Engineering

基　　金：国家部委基金资助项目(9140A27040413JB11407)

摘　　要：针对现有的多策略本体映射方法大多采用固定的组合方式以及存在映射效率不高的问题,提出了一种自适应的多策略本体映射方法。该方法首先从名称、语言、实例和结构4个维度定义了4种不同的本体映射策略;然后,对每种策略分别提取中间映射结果,并利用优化的数据结构进行存储。通过计算各个策略的局部置信度,实现对中间映射结果的动态组合;最后,利用启发式贪心算法提取映射结果。实验表明:该方法不仅能够提高映射效率,而且取得了更好的映射效果。In view of the existing multi-strategy ontology matching methods which mostly use a fixed composition style,and have the problem of low matching efficiency,this paper proposes an adaptive multi-strategy ontology matching method.Firstly,this method defines four different strategies for ontology matching from four dimensions,i.e.,name,language,structure and instance,extracts mappings from each strategy respectively,and stores the intermediate mapping results by using an optimized data structure.Through the calculation of local confidence of each strategy,the dynamic composition between strategies is realized.Finally,by using heuristic greedy extract algorithm,the mapping results are obtained.Experimental results show that this method can not only improve the matching efficiency,but also achieve better mapping results.

关 键 词：本体映射 多策略组合 局部置信度 映射效率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]