Analysis Of Auto Insurance Premium Pricing Based On GAMLSS Model

The auto insurance industry has developed rapidly,and the demand for auto insurance has also been increased.This makes motor vehicle premiums continue to occupy the first place in property insurance premiums.Maintaining the stability benefits and health status of motor vehicle insurance is the core content of the state and property insurance companies.However,the existing method of determining motor vehicle premiums is monotonous and cannot provide customers with personalized products.The reform of commercial auto insurance rates is in full swing.This requires insurance companies to be more scientific and refined in determining auto insurance rates to meet the human needs of customers.Therefore,we should innovate products and services.And we should innovate premium determination and calculation methods.So that premiums and risks are more matched,and we will find a better method for determining premiums.This article explores auto insurance premiums individually,predicts the customer's claim amount,and uses this as a basis for premium pricing.First establish a pure premium one-stage model and a pure premium two-stage model.Compare the advantages and disadvantages of the two models.Two generalized linear models,generalized Poisson regression and generalized Gamma regression,are established based on the number of claims and the average claims in the two-stage model of pure premiums.Based on the generalized linear model(GLM),we consider fully expanding the GLM model.And a generalized additive model based on position,scale and shape(GAMLSS)is established.In the end,the negative binomial type II distribution GAMLSS model is established for the number of claims,and the inverse Gaussian distribution GAMLSS model is established for the average compensation.The fitting effect is better.The research focus of this paper is to compare the fitting effects of different models through pure premium forecasting analysis.The aim is to quarry the optimal model.By continuously adjusting the model,the model is more applicable.The auto insurance pricing and calculation methods are enriched,and personalized insurance pricing is achieved.