基于部分标准对齐的协同遗传规划算法解决本体匹配问题  

Solving ontology matching through collaborative genetic programming algorithm based on partial standard alignment

作　　者：姜照航 吕青 戴可涛 孙东磊 Jiang Zhaohang;Lyu Qing;Dai Ketao;Sun Donglei(College of Electrical&Power Engineering,Taiyuan University of Technology,Taiyuan 030024,China)

机构地区：[1]太原理工大学电气与动力工程学院,太原030024

出　　处：《计算机应用研究》2025年第4期1095-1101,共7页Application Research of Computers

基　　金：山西省省筹资金资助回国留学人员科研资助项目(2023061)。

摘　　要：本体异质性阻碍了知识交互和数据共享。本体匹配通过整合相似度特征解决多种异质问题。为提高匹配结果的质量,提出了基于部分标准对齐(PSA)的协同遗传规划算法。该算法利用基于自适应概率的交叉策略和基于语义相似度的变异策略,平衡局部开发和全局探索,寻找高质量的相似度表示;其次,提出了一种新的适应度函数,引导算法的寻优方向,全面挖掘PSA中的信息价值;最后,提出了一种新的PSA修正方法,通过多维度识别可疑匹配对,并结合随机森林分类器聚合专家投票,优化PSA来构建更可靠的相似度特征。实验结果表明,该方法在不同的专家错误率和本体匹配任务中均能获得高质量的匹配结果,优于先进的匹配技术。Heterogeneity in ontologies hinders knowledge interaction and data sharing,while ontology matching addresses various heterogeneity issues by integrating similarity features.To improve the quality of matching results,this paper proposed a collaborative genetic programming algorithm based on partial standard alignment(PSA).The algorithm employed an adaptive probability-based crossover strategy and a semantic similarity-based mutation strategy to balance local exploitation and global exploration,aiming to discover high-quality similarity representations.Additionally,a new fitness function guided the optimization direction of the algorithm,fully leveraging the informational value of PSA.Finally,a novel PSA correction method identified suspicious matching pairs through multiple dimensions and aggregated expert votes using a random forest classifier,optimizing the PSA to construct more reliable similarity features.Experimental results demonstrate that the proposed method achieves high-quality matching results across different expert error rates and ontology matching tasks,outperforming state-of-the-art matching techniques.

关 键 词：本体匹配 部分标准对齐 协同遗传规划 适应度函数 PSA修正 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]