Research On Vehicle Risk Management And Control From The Perspective Of Automobile Insurance Operation

Since the reform and opening up,with the popularization of automobiles and the improvement of people’s safety awareness,people have paid more and more attention to the problem of vehicle risk management and control.However,risk cases such as vehicle risk assessment,vehicle risk identification,and vehicle risk management have occurred from time to time.At present,most scholars at home and abroad have studied a series of issues such as vehicle risk levels and vehicle risk management and control.However,China’s vehicle risk management started late and has not yet formed an effective and complete system.Therefore,the study of vehicle risk control has great theoretical and practical value.This paper focuses on the underwriting and claims data of a property insurance company from 2014 to 2016,and uses the analytic hierarchy process,data mining techniques and game theory to study the network system of “insurant-insurer-maintenance unit”.First,it analyzes vehicle risk management issues such as vehicle risk assessment under the perspective of policyholders,vehicle risk identification under the perspective of insurers,and vehicle risk management under the perspective of maintenance units;then,the vehicle risk control model is built and optimized based on the network composed of “insurant-insurer-maintenance unit”.The specific research contents are as follows:(1)In order to minimize the cost of vehicle operations,the vehicle risk assessment is studied from the perspective of policyholders.The risk factors are divided into four types of influencing factors,which are driver factors,vehicle factors,driving factors,regional environmental factors,and the risk factor options of each influencing factor are scored.Then,the analytic hierarchy process is used to determine the weight of each index,and the scientific effectiveness of the vehicle risk assessment method is verified by comparing with the low risk customer data that has been insured by the insurance company.(2)In order to minimize the operating costs of vehicles,the risk identification of vehicles is studied from the perspective of insurers.Taking the vehicle insurance claim cases with a loss of more than 2,000 yuan of an insurance company for 2014-2016 years as the research objects,the data is divided into two groups: the experimental group and the verification group.Taking the influencing factors of the amount of auto insurance loss as the input variable and taking the risk level as the output variable,the BP neural network prediction model is constructed using the experimental group data.Then,the model is used to predict the verification group data and the rationality of the model is verified.Then,taking the 2014-2016 vehicle fraud case of the insurance company as the research object,the risk factors of the car insurance fraud(accident liability,time of departure,whether to report the crime at the scene,etc.)are statistically extracted and classified.The Apriori algorithm in the association rules is used to find a high frequency fraud risk factor combination with strong correlation between minimum support threshold and minimum confidence threshold in order to provide the basis for the vehicle risk management decision of the insurer.(3)In order to minimize the cost of vehicle maintenance and improve the quality of maintenance services,based on the classification of vehicle accessories market,the risk analysis is carried out from the perspective of the price and quality of vehicle accessories.From the perspective of maintenance man-hour cost,the factors that lead to the increase of vehicle maintenance man-hour risk are analyzed.In addition,the GM(1,1)Markov prediction model is constructed by using the repair amount accuracy data of the auto insurance claims of a property insurance company in 2017 as the original data,and then the repair amount precision of the auto insurance claims of the insurance company in the 1-3 month of 2018 is predicted.(4)In order to maximally balance the interests and risk control of each participant in the vehicle risk network,a dynamic game model of three types of participants’ network interest equilibrium control is constructed,including policyholders,insurers and maintenance units.Taking the above research results as the data basis,the projection algorithm is constructed to analyze and solve the problem,determine the rule of influence of each behavior decision on vehicle risk sharing,and tap the equilibrium point of network interests.