Research On Semantic Retrieval Of Civil Aviation Unsafe Events Based On Ontology

In recent years,China’s civil aviation safety management has achieved substantial results due to the continuous improvement of the security management level.Information generated by the civil aviation unsafe events is a main resource on which the security management activities depend.High-quality information retrieval,analysis and prediction of unsafe events have advanced and preventive significance.The traditional retrieval based on keyword matching is just the matching method at the level of words,which cannot understand the information at the semantic level and leads to the fact that query results deviate from the actual needs of the users.However,semantic retrieval provides a kind of method to realize information retrieval from the semantic level based on semantic understanding and using the association between information resources,which is a hot research direction of information retrieval technology.This paper studies and analyzes the related knowledge of ontology and information retrieval technology on the basis of current situation of civil aviation domain knowledge,which includes the concept,classification,construction method,semantic relations and description language of ontology and the comparison of the traditional keyword matching method and semantic retrieval method.Afterwards,aiming at the problem of the lack of unified knowledge model in the field of civil aviation unsafe events,the domain ontology of unsafe events is compiled using the tool of Protége with the methods of Stanford University seven steps and continuous efforts and improvements.The process involves the acquisition of relevant knowledge of the field,the establishment of the concept hierarchy structure,the addition of the concept property structure,encoding and formalization.And then use the ontology description language to describe the civil aviation unsafe events ontology.On the basis of the constructed unsafe events ontology,the semantic retrieval method is researched and the semantic retrieval process is put forward.At first,the method of semantic annotation in the progress of semantic retrieval is studied.Then,an improved semantic similarity algorithm based on semantic distance and ontology concept information is proposed,which is also applied to the computation of the semantic vector weights of the unsafe events documents.What’s more,the method of semantic query expansion based on the ontology is discussed.Finally,semantic retrieval is achieved and verified through an example.The result demonstrate that semantic retrieval based on unsafe events ontology can effectively reduce the problem that query results deviate from the actual needs of the users caused by traditional retrieval method and improve the recall and precision of the information retrieval.The research of this paper is still in a stage of relatively basic exploration,and there are some problems and shortcomings.In future work,the semantic reasoning tools such as JENA and the related development environment will be applied to realize the real semantic retrieval based on the unsafe events ontology.Ideally,it could be used in the actual work to facilitate users to study and utilize.