Research On Integration Method Of Sensor Ontology Matching Results

The equipment data of each system in the power internet of things has the characteristics of multi-source and highly heterogeneous,which makes it difficult to interact and share among them.Sensor ontology is the core of semantic technology,providing effective data and knowledge representation,management and sharing technical means for collaboration between sensor data.To solve the multi-source heterogeneous problem,it is necessary to convert the equipment data of each system into a sensor ontology and integrate the heterogeneous data entities in different sensor ontologies.However,because there is no similarity measure that can ensure that heterogeneous sensor data entities can be effectively identified under all circumstances,it is necessary to comprehensively consider multiple similarity measures to ensure the quality of the matching results.How to effectively integrate the matching results of sensor ontology determined by different similarity measures to ensure the quality of the final matching results is a research hotspot in the field of power internet of things ontology.Around this scientific problem,scholars at home and abroad have carried out a lot of research in recent years.The related work can be roughly divided into the following two categories:(1)Solving the problem of extracting the matching results of sensor ontology,That is,how to directly extract high-quality matching targets from the matching results determined by different similarity measures;(2)Solving the meta-matching problem of sensor ontology,that is,how to assign the optimal integration weight to the matching results determined by different similarity measures to improve the integrated matching results quality.This paper carries out the following research work in response to the above two problems:(1)Aiming at the problem of extracting the matching results of sensor ontology,facing the characteristics of sensor concepts with rich semantics and complex relationships between concepts,based on the traditional serial and parallel integration methods,“a method for extracting matching results of sensor ontology based on the hybrid semantic similarity measure” is proposed.This method first treats the extraction problem as a binary classification problem,and then uses the “parallel-serial” hybrid integration method of the similarity measures to extract high-quality matching results.(2)Aiming at the meta-matching problem of sensor ontology,facing the large scale of candidate integration weights,based on the construction of three types of ontology heterogeneity metrics,“a sensor ontology meta-matching method based on ontology heterogeneity metric” is proposed.This method first regards the meta-matching problem as a regression problem of “ontology heterogeneity value-integrated weight”,and then fits the regression function of “ontology heterogeneity value-integration weight” by extracting a representative entity set to adaptively determine the integrated weights of the similarity measures.The test data set used in this paper is the data set provided by the International Ontology Alignment Evaluation Initiative(OAEI)and two pairs of real sensor ontologies.First,this paper compares “a method for extracting matching results of sensor ontology based on the hybrid semantic similarity measure” with the measures of OAEI participants and heuristic-based matching methods.The experimental results show that this method can effectively improve the quality of the sensor ontology matching result.Then,this paper compares “a sensor ontology meta-matching method based on ontology heterogeneity metric” with OAEI participants.The experimental results show that the method can adaptively determine the effective integration weight of the matching result according to the heterogeneous characteristics of the sensor ontologies to be matched,thereby ensuring the quality of the matching result of the sensor ontology.The research work in this article can provide new research directions and research ideas for the design of the integration method of sensor ontology matching results.The related research results can be used to integrate all kinds of heterogeneous sensor ontologies,promote the realization of interaction and sharing of device data in power Internet of things,and lay the foundation for display service,advanced data service,intelligent application service of power equipment.