An approach for linking dynamic network information models based on ontology matching
Igor Kulikov, Jiafeng Yang, Nataly Alexandrovna Zhukova, Man Tianxing · Expert Systems with Applications · 2025
• Novel context-aware ontology matching approach that links heterogeneous dynamic network models using business process context, eliminating the need to replace existing proprietary information systems while enabling standardized interoperability. • Dynamic maintenance mechanism that automatically updates ontology links when source models change, requiring updates only when new element types are added or deleted, ensuring predictable and manageable system evolution. • The Integration framework achieves F-measure scores greater than 0.72 across multiple duplicate detection algorithms, with successful demonstration linking cable television operator networks to TM Forum standardized models. Dynamic network information models are typically heterogeneous and isolated systems that impede effective interoperability, significantly hindering end-to-end service integration and data sharing across network segments. To address this challenge, we propose a new approach for linking heterogeneous dynamic network models based on ontology matching, which can be applied in various domains utilizing dynamic networks. For ontologies matching we use different existing duplicate detection algorithms but we reduce the computational complexity of ontology matching due to splitting initial set of matched entities into a number of subsets using domain knowledge. Using telecommunications as case study, we represent operator networks as knowledge graphs and match them with standardized model ontologies using business process context to create an Extended Operator Network Ontology. Our approach ensures linking of dynamic network models used in operators information systems that is of primary importance for implementing complex business processes, and providing integrated services while maintaining existing models.