Research On The Contributing Factor To The Loss Ratio Of Auto Insurance In XX Property Insurance Branch Based On Random Forest Algorithm

With the rapid development of China’s economy and the gradual improvement of national living standards,the demand for insurance products in people’s daily production and life is gradually increasing.Considering the increasingly weak development of China’s motor vehicle market and the decreasing share of auto insurance premiums in property and casualty insurance companies,China officially promulgated the "Guidance on the Implementation of Comprehensive Reform of Auto Insurance" on September 18,2020,star ting a new round of auto insurance comprehensive reform.Now that it has been more than one year since the 2020 fee reform,the national auto insurance is facing the current situation of double rise of comprehensive claim rate and comprehensive cost rate,therefore,it is of good practical significance to study the contributing factor to the loss ratio of auto insurance.This paper first provides an overview of the current scholars’ influencing factors and research methods on auto insurance loss ratio based on literature research.Then the concepts and theoretical foundations closely related to this paper are introduced,followed by an analysis of the reform history of the auto insurance market and the current situation of the national auto insurance operation,and the real data of XX General Insurance Branch are selected to analyze the current situation of its loss ratio.In order to further study the factors affecting the loss ratio,this paper discusses the influence of auto insurance loss ratio from four aspects: subject matter characteristics and driver characteristics,insurance underwriting and claims,third party partners,and social and natural environment,based on the existing literature.Then,we compiled data of more than 680,000 policies underwritten by XX General Insurance Branch after the auto insurance reform,selected 17 feature factors,and applied the random forest model in the field of machine learning to predict whether the policy was a high payout risk policy.The results of the study show that the most influential factor among the selected factors on the explanatory variables is the characteristics of the insurance claims process,so the insurance companies focus on the cost control of the actual occurrence of complex and heavy cases;the insurance coverage,driver’s age,the number of car insurance policies,the age of the car,the nature of the vehicle and other characteristics of the insurance subject and driver characteristics have a greater impact on the policy’s high claims;whether the high speed,whether the time period of the insurance is 9 pm to 11 pm Other noteworthy factors are whether the time of insurance is close to the policy termination period,and insurance companies need to pay attention to the authenticity of the accident during this special period and whether it contains the risk of duplicate claims.Finally,taking into account the results of the empirical analysis and the actual operation situation of XX General Insurance Branch,we propose targeted countermeasures in three aspects,namely,improving the accurate pricing ability,optimizing the claims operation cost,and strengthening the anti-fraud work,to help enhance the company’s operation efficiency and improve the loss ratio of auto insurance.