Research On Knowledge Discovery Of Railway Risk Inter-relation Based On Ontology

In the traditional the research of heavy rail transportation risk research often focus on the linear causal relationship between risk and accident,however the mechanism of accident causing is the complicated situation of mutual influence,the risk of polymerization,the final form of the upgrade in reality.In order to effectively reduce the accident frequency,pre-control of risk source,it is necessary to study on association mode between source risk characteristics.Based on the characteristics of accident in heavy-haul railway,this paper combines text mining and data mining,ontology-based knowledge discovery and knowledge reasoning to study the characteristics of risk source correlation,and provides accurate prediction for risk pre-control.The main research contents are as follows:(1)Ontology construction of railway risk:Based on the accident report of the years,the paper describes the cause of different accidents.By analyzing the causal mechanism of the several of risk sources(environment,equipment,personnel,etc.)in different accident scenarios,risk sources of knowledge reuse for the purpose of constructing the initial risk model of the heavy haul railway.(2)Railway risk analysis:According to the risk database of Heavy Haul Railways(semi structure data and text data),extract the key words of risk source from semi-structured data by text analysis method,which may affect the risk factors of the accident,and verify the influence of risk factors on the accidents.Adding risk factors to the heavy haul risk ontology.(3)Data mining based on risk data:Because the upgrading of railway accidents has the characteristics of delay,based on the traditional "Apriori" association rule algorithm,research add the ability of time series analysis,using the improved algorithm of risk data of heavy haul railway,mining the correlation model between risk factors,further analysis of the underlying railway accident caused mechanism,adjust the expansion the risk of heavy haul railway accident analysis based on Ontology model.(4)Ontology generation of railway risk:Propose a domain ontology for heavy haul railway heavy haul railway risk data semi-automatic construction method of ontology construction,ontology mapping,through collecting data and code to generate the description ontology check.Semi-automatic construction method,which combines the efficiency of ontology knowledge reasoning and the reliability and domain knowledge of ontology,and provides a feasible method for ontology learning based on structured and semi-structured data sources.(5)Ontology learning of railway risk:Based on knowledge of risk association,the paper study the knowledge reasoning and updating of risk ontology of heavy haul railway,and the knowledge modeling of heavy haul railway based on ontology is put forward.The knowledge model is constructed by knowledge layer structure analysis and conceptual model of ontology layer.And the application of knowledge of heavy railway is promoted,and the early warning accuracy of accident and the efficiency of dealing with the accident are realized.At the same time,railway risk knowledge sharing and reuse are realized,and the rapid acquisition and maintenance process of railway risk knowledge.