| The introduction of the concept of smart transportation provides new ideas and directions for the development of urban public transportation.In January 2022,the State Council issued the "14th Five-Year Plan for the Development of a Modern Comprehensive Transportation System",requiring "an orderly construction of an intelligent management platform for urban transportation and strengthening the refined management of urban transportation." However,the diverse sources of urban public transportation data have brought great challenges to data management and application.How to clearly define these massive data is crucial to the realization of data semantic interoperability between different information systems,and it is an urgent problem to be solved.As a formal representation method for describing and organizing concepts,entities and their relationships,ontology can integrate and share data from different systems,thus supporting cross-departmental and cross-domain data applications.However,there are differences in concepts,attributes,and relationships among the ontologies of data from different sources,and the data cannot be merged and used directly based on heterogeneous ontologies,which affects the sharing and integration of information.Ontology mapping can match and map the concepts and semantics of heterogeneous ontologies to achieve semantic consistency of data.However,the existing ontology mapping methods have problems such as low accuracy of similarity calculation results and poor mapping effects,and there is no relatively comprehensive or complete ontology that can be reused in the field of urban public transportation services.This thesis studies the ontology similarity calculation method based on graph attention network,and based on this method,proposes an ontology mapping method based on instance feedback,constructs an urban public transport service ontology,and realizes the association of multi-source urban public transport service data.Finally,an ontology mapping system for urban public transport services is designed and developed based on the ontology mapping method.The main research work is as follows:(1)Ontology similarity calculation is an important basis for ontology mapping,but the existing ontology similarity calculation methods are less accurate.Aiming at this problem,this thesis proposes a graph attention network-based method for computing the similarity of heterogeneous ontologies.This method constructs the ontology as an undirected topological graph,and introduces an attention mechanism to dynamically consider the influence of edge weights to achieve better neighbor aggregation.At the same time,it is more robust to noisy neighbors and can achieve higher calculation accuracy.(2)Ontology Mapping Semantic mapping of multi-source ontology to realize integration and interactive operation between different ontologies.However,the current ontology mapping methods mainly rely on similarity calculations,and the mapping results are not good.This thesis proposes an ontology mapping method based on instance feedback.On the basis of using graph attention network model to calculate ontology similarity results,a preliminary mapping is formed,and then instances are used as feedback to adjust the mapping rules.Experiments on ontology mapping of urban public transport services demonstrate the effectiveness of the method.(3)Urban public transport service is a field involving many stakeholders,and improving service efficiency and level requires analyzing the required data from different scenarios.This thesis adopts the internationally common enterprise modeling method(TOVE),starts from the required capabilities,analyzes the data types required in different scenarios,builds a unified urban public transport service ontology,and provides mapping for the ontology of multi-source urban public transport service data Benchmark,addressing ontology heterogeneity.(4)Based on the ontology mapping method proposed in this thesis,design and develop a set of ontology mapping system for urban public transport services.Users can realize the mapping of heterogeneous ontology and information query based on ontology,and visualize the results.In summary,this thesis proposes a ontology similarity calculation method based on graph attention networks,and introduces instance feedback to improve mapping rules in ontology mapping,which enhances the accuracy of ontology mapping.A city public transportation service ontology is constructed and applied to the field of urban public transportation services,addressing the issue of heterogeneous city public transportation service ontologies from multiple sources and supporting the integration and semantic manipulation of city public transportation service data. |
