Risk Assessment Of Economic Loss In Debris Flow Disasters Based On Bayesian Network

In recent years,natural disasters occur frequently around the world.China is a country,which is most seriously affected by natural disasters.Yunnan province is located in the southwest of China,with special geographical and natural environment.It is a province where natural disasters occur frequently.Due to its location at the junction of mountains and plates,the terrain is undulating and the climatic conditions are special.Because the debris flow disaster has strong destruction,the property of the facilities will be severely damaged.There have been several serious debris flow disasters,and a large number of economic losses.In urgent need of further study of the debris flow disaster risk to reduce the influence of the local social and economic development.In the process of disaster risk management,it is an important step to analyze and evaluate disaster risk.According to the existing research results,the occurrence of debris flow disaster economic loss is related to the disaster itself,and associated with human society,the process of debris flow disaster risk assessment should be comprehensive and considered the risk of debris flow disasters and the vulnerability of human society.Because of its strong ability to analyze uncertain problems,bayesian network model can not only integrate multiple factors to complete the evaluation,but also can show the logical relationship among factors in a graphical way,which has been widely used in disaster risk management.Based on the basic characteristics of debris flow,this article firstly combining with the relevant data to construct a bayesian network by comparing with ANN and SVM,based on the network structure,the logical relationship between debris flow hazard factors showed,and use the network model to realize the dangers of debris flow disaster prediction.Secondly,the POT model with threshold 480 is used to obtain the probability distribution of the economic loss of local debris flow disaster.Combining the bayesian network prediction results and the probability distribution of economic loss,the VaR and ES of each region are calculated by selecting different confidence levels,so as to realize the monetization of the risk of disaster economic loss.Finally,the results providing suggestion and reference for future disaster risk management.