Research On Automobile Insurance Rate Determination Based On Double Hierarchical Generalized Linear Model

In recent years,with the development of China’s insurance industry gradually getting on the right track,people’s demand for auto insurance has become higher and higher,and the proportion of auto insurance in non-life insurance has always been more than 65%.Due to the late start of the research on the rate determination of domestic automobile insurance,the common rate determination methods have not kept up with the demand of the existing domestic market.How to further improve the actuarial method of automobile insurance and scientifically determine the rate is of great significance to promote the healthy and high-quality development of China’s automobile insurance market.The generalized linear model(GLM)is the most widely used model in the rate determination of automobile insurance.Its advantages of low data requirements and loose model setting have advantages in the determination of vehicle insurance rates,overcoming the defects of the classical linear regression model.However,in practical applications,GLM still has limitations in its application.If the discussion of random effects is not sufficient,GLM will have a large deviation when processing large amounts of data.The double hierarchical generalized linear model(DHGLM),as an extended model of GLM,provides a new idea for the method of determining vehicle insurance rates.Because DHGLM does not require that the model parameters must be fixed constant values,and combines the advantages of the generalized mixed model(GLMM)and the hierarchical generalized linear model(HGLM),it introduces random effects,relaxes the assumption that the variables obey the distribution requirements,and improves the applicability of the model by establishing GLM for the sub-model and the main model respectively.This paper uses DHGLM for analysis and modeling.Based on the real vehicle insurance claim data,it compares the results of GLM’s vehicle insurance rate determination from three aspects: parameter significance,data interpretation of the model and the amount of pure premium.The results show that compared with GLM,DHGLM is more reliable and widely used in the application of vehicle insurance rate determination.At the end of the paper,according to the empirical conclusions,some relevant countermeasures and suggestions are also put forward.The main innovations of this paper are as follows: 1.The introduction of DHGLM into the study of automobile insurance rate determination is an innovation of the actuarial model of automobile insurance rate determination.2.The iterative general term formula of the new DHGLM sub-model is proposed,which improves the efficiency of the model and improves the application of the model in the field of vehicle insurance rate determination.