Accident Causes Knowledge Graph Construction Based On Chemical Cases

In recent years,fire,explosion,poisoning and other chemical accidents occur frequently,causing serious loss of life and property,which has attracted wide attention from academic community.Accident prevention,as a means of risk management that can effectively avoid accidents,can find the rules of accidents by analyzing the causes of past accidents,so as to put forward improvement plans to prevent similar accidents from happening again.However,due to the limited text form of accident cases used for analysis at present,it is difficult to support the analysis of the causes of the accident cases Therefore,it is urgent to find a way to structure text data to support the formulation of prevention strategies and regulations.As a special database that integrates information into ontology,knowledge graph provides ontology support and technical method for textual data structuring.In order to build the knowledge graph,accident investigation reports were first collected from the safety management network and government emergency management network.According to the characteristics of the collected data,the following three aspects were studied in this paper:(1)Construction of knowledge graph pattern layer.In this paper,the reason’s classification of the 2-4 accident model was adjusted based on text data,and the elements of causes in the accident were determined by combining with the SEM event representation model,thus,the pattern layer ontology model of the knowledge graph was constructed.(2)Construction of Knowledge Graph Data Layer.According to the process of reason extraction: cause identification,cause element identification,and cause relation identification,the causes,entities and causality in the knowledge graph were extracted in this paper,thus,the data layer of the knowledge graph was constructed.(3)Refinement of knowledge graph pattern layer.In this paper,the cause factors were obtained by clustering the cause cases in the data layer,and the accident cause classification system was constructed manually by combining the expert knowledge.Then,the Apriori algorithm was used to mine the correlation between the cause factors.Finally,the knowledge in the knowledge graph was imported into the database in the form of triples.Thus,the construction of chemical accident cause knowledge graph was completed.(4)Finally,in this paper,some preventive suggestions were put forward for chemical accidents based on knowledge graph.Based on the DEMATEL influence analysis method,the correlation between the cause factors was analyzed in this paper,so as to obtain the key factors that have a greater impact on the system,and puts forward corresponding suggestions for the key cause factors.Combined with the DEMATEL analysis method,in this paper,the knowledge graph finally constructed was analyzed.The key elements of the causative system were found out,and the corresponding suggestions were put forward to provide some reference for the prevention of similar chemical accidents.