采用多目标粒子群算法的本体元匹配方法

Ontology Meta-Matching Method based on Multi-Objective Particle Swarm Optimization

为了解决本体之间存在的异构问题,提出一种本体元匹配方法来确定不同本体中实体之间的对应关系.首先设计两个本体匹配结果质量的近似度量方法,并在此基础上构建本体匹配问题的多目标优化模型,最后提出一种多目标粒子群算法以求解该问题并优化本体匹配结果的质量.采用国际本体匹配竞赛提供的benchmark测试集,来测试基于多目标粒子群算法的本体元匹配方法的性能.实验结果表明:提出的方法在查全率和查准率两个指标上均优于国际本体匹配竞赛的其他参与者.

To address the problem of heterogeneity among different ontologies,this paper proposes an ontology meta-matching method to determine the corresponding relationship between entities in different ontologies.firstly,two approximate measurement methods for the quality of ontology matching results are designed.Then,a multi-objective optimization model of ontology matching problem is constructed on this basis.Finally,a multi-objective particle swarm optimization is proposed to solve the problem and optimize the quality of ontology matching results.On the other hand,the experiments in this paper use the benchmark test set provided by the International Ontology Alignment Evaluation Initiative to test the performance of the ontology meta-matching method based on multi-objective particle swarm optimization.Generally speaking,the experimental results show that the proposed method outperforms the other participants of Ontology Alignment Evaluation Initiative in both recall and precision.