Proceedings of the
The Nineteenth International Conference on Computational Intelligence and Security (CIS 2023)
December 1 – 4, 2023, Haikou, China
Efficient Ontology Matching through Co-Evolutionary Compact Genetic Algorithm with Niching Strategy
1Fujian Provincial Key Laboratory of Big Data Mining and Applications, Fujian University of Technology, Fuzhou, Fujian, China.
2School of Computer Science and Mathematics, Fujian University of Technology, Fuzhou, Fujian, China.
3Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing, Guilin Universitiy of Electronic Technology, Guilin, Guangxi, China.
4Library, Fujian University of Technology, Fuzhou, Fujian, China.
ABSTRACT
Ontologies provide a standardized format for knowledge representation, enabling its distribution across different domains. However, interoperability issues can arise due to the use of different ontologies or vocabularies by distinct systems and partners. By leveraging heterogeneous ontology alignment, intelligent applications can proficiently exchange information, enhancing communication and decision-making while minimizing data integration costs. In this study, we delve into the challenge of matching ontologies, with the goal of identifying an optimal set of concept pairs with the highe f-measure value. Due to the complexity inherent to ontology matching, we resort to a Genetic Algorithm (GA) as our solution framework. In this approach, we first depict the Heterogeneous Ontology Matching problem (HOMP) as a multi-modal issue with sparse solutions, and then we propose a Co-Evolutionary Compact Genetic Algorithm with Niching Strategy (CCGA-NS) to address it. CCGA-NS employs probability distribution estimation to simplify population representation, following which it maintains three separate virtual populations with different evolutionary strategies in search of the global optimum. Our experiment includes test cases from the OAEI Conference track. The results demonstrate that CCGA-NS surpasses other leading ontology matching methods in terms of efficiency and performance.
Keywords: Ontology matching, Compact genetic algorithm, Co-evolutionary mechanism, Niching strategy.
Download PDF
1Fujian Provincial Key Laboratory of Big Data Mining and Applications, Fujian University of Technology, Fuzhou, Fujian, China.
2School of Computer Science and Mathematics, Fujian University of Technology, Fuzhou, Fujian, China.
3Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing, Guilin Universitiy of Electronic Technology, Guilin, Guangxi, China.
4Library, Fujian University of Technology, Fuzhou, Fujian, China.
ABSTRACT
Ontologies provide a standardized format for knowledge representation, enabling its distribution across different domains. However, interoperability issues can arise due to the use of different ontologies or vocabularies by distinct systems and partners. By leveraging heterogeneous ontology alignment, intelligent applications can proficiently exchange information, enhancing communication and decision-making while minimizing data integration costs. In this study, we delve into the challenge of matching ontologies, with the goal of identifying an optimal set of concept pairs with the highe f-measure value. Due to the complexity inherent to ontology matching, we resort to a Genetic Algorithm (GA) as our solution framework. In this approach, we first depict the Heterogeneous Ontology Matching problem (HOMP) as a multi-modal issue with sparse solutions, and then we propose a Co-Evolutionary Compact Genetic Algorithm with Niching Strategy (CCGA-NS) to address it. CCGA-NS employs probability distribution estimation to simplify population representation, following which it maintains three separate virtual populations with different evolutionary strategies in search of the global optimum. Our experiment includes test cases from the OAEI Conference track. The results demonstrate that CCGA-NS surpasses other leading ontology matching methods in terms of efficiency and performance.
Keywords: Ontology matching, Compact genetic algorithm, Co-evolutionary mechanism, Niching strategy.
Download PDF