Research On The Construction Method Of Event Rational Graph Based On Common Chemical Safety Accident Investigation Reports

In recent years,with the continuous development of chemical technology,the frequency of chemical safety accidents is also rising.As a pillar industry of national economic development,chemical enterprises will not only cause personal injury,lead to significant property losses,but also affect social stability and damage to the ecological environment.The analysis of past chemical accident cases can prevent the occurrence of accidents,improve the safety production level of chemical enterprises,enhance the emergency rescue ability of chemical accidents,and effectively prevent and reduce major accidents in chemical production.The knowledge graph can help realize the unified representation and management of data in the field of chemical safety,and assist in accident prevention through intelligent analysis.Therefore,based on the investigation report data of common chemical safety accidents,this paper analyzes,extracts and constructs entities and relationships,connects relevant domain knowledge,and constructs the event process graph in the field of chemical safety.The main work content is as follows:(1)Construct a data set in the field of chemical safety.Due to the lack of data sets in the field of chemical safety,this paper collected nearly a thousand original investigation reports of chemical accidents in the safety management network,the government emergency management network and other websites,and screened,cleaned and labeled the text data of the reports to construct a data set in the field of chemical safety.(2)Constructing the event and principle graph in the field of chemical safety.Based on the general accident cause model and the characteristics of chemical accident investigation report text corpus,this paper constructs the ontology model of the graph.On this basis,the rule-based method is used to extract knowledge from semi-structured data.The Bi LSTM-CRF model is used to extract knowledge from unstructured data,and multiple sets of comparative experiments are set up to verify the effectiveness of the model.Then,the template matching method based on causal prompt words is used to extract the causes and effects of the accident.Finally,through knowledge fusion,the extracted information is stored in the Neo4 j graph database in the form of triples to complete the construction of the event rational graph.(3)Design and implementation of automatic question answering system based on event reasoning graph of common chemical safety accident investigation report.According to the system requirements,a Web system interface based on Django framework is built,and the knowledge retrieval and acquisition of the event process graph is realized based on the custom question and answer template and Cypher query sentence,which verifies the effectiveness and practicability of the event process graph based on common chemical safety accident investigation reports constructed in this paper.