Research On Emergency Process Reasoning Based On Risk Cascade Knowledge Graph

Construction supply chain risk has increasingly become a key issue affecting the profitability of construction supply chain enterprises.Taking a large domestic construction supply chain enterprise(hereinafter referred to as "W enterprise")as the research object,based on risk management theory and knowledge graph theory and collaborative scenario modeling theory and other theories,this paper has successively built multi-layer network mapping model,risk cascade knowledge graph and emergency process knowledge base from the perspective of post emergency.On this basis,the emergency process reasoning algorithm has been designed and applied,and the prototype of emergency information system has been developed.This paper aims to help W enterprise reduce the loss caused by risk events in time through emergency process reasoning.The specific research contents of this paper are as follows:(1)Construction of W enterprise risk cascade knowledge graph: according to the procurement service content and risk documents of W enterprise,this paper sorts out the risks and risk events in a number of business activities including procurement services from the organizational and business levels.Based on the risk correlation results,the multi-layer network mapping model and W enterprise risk cascade knowledge graph 1.0and 2.0 are constructed.(2)Construction of W enterprise emergency process knowledge base: emergency process knowledge is extracted from the risk list,internal control manual and practical experience of W enterprise,and sorted into structured text data,and then imported into the multi-layer network mapping model to form emergency process layer.Based on this,the extracted and organized emergency process examples are added to the risk cascade knowledge graph of W enterprise,which means that the emergency process knowledge is stored in the form of graph data and the emergency process knowledge base of W enterprise is established.(3)Design and application of emergency process reasoning algorithm: based on the emergency process knowledge base of W enterprise,combined with the characteristics of risk events,the emergency process reasoning algorithm is designed to efficiently match the emergency process.In order to achieve this aim,emergency process reasoning algorithm integrates function,text mining algorithm and machine learning algorithm,and the effectiveness of its reasoning emergency process is tested by case application.According to the results of reasoning algorithm,the mapping relationship between risk event layer and emergency process layer is added to the multi-layer network mapping model.Then match the corresponding emergency process examples for the risk event examples in the risk cascade knowledge graph and finally build a W enterprise risk cascade knowledge graph 3.0 that can characterize the knowledge reasoning logic and risk cascade effect of "Organization-Business-Risk-Risk event-Emergency process".At the same time,the application effect of emergency process reasoning algorithm is evaluated from the two dimensions of reasoning efficiency and derived risk.(4)Development of prototype of emergency information system: based on the emergency process reasoning algorithm,the prototype of emergency information system is developed,and the concept of collaborative scenario engine is introduced to meet the actual business needs of W enterprise to the greatest extent.Specifically,the multi-layer network mapping model,risk cascade knowledge graph and emergency process reasoning algorithm obtained above are integrated into the emergency information system prototype to enhance the landing of research results and give better play to its ability in dealing with risk events.In the test of the prototype system,the visualization process of emergency process reasoning is shown.At the end of this paper,it summarizes the research content in the sense of management and looks forward to the future of relevant research based on the shortcomings.