Application And Research On Zero Inflation Model In The Number Of Auto Insurance Claims

Motor vehicle insurance, car insurance for short, is the biggest insurance innon-life insurance business, its development foreground is very wide. To solve theproblem of classifcation ratemaking, generalized linear model is most commonlyused in ratemaking for motor vehicle insurance. But when claim frequency dataappear a lot of zero, namely zero-infated, generalized linear model is no longer ap-plicable, in this case, zero-infated models could solve this situation. In zero-infatedmodels, we generally assume that the proportion of structural zeros is a constantand not infuenced by rate factors, but this may deviate from the actual situation.This paper assumes that the proportion of structural zeros is not a constant andafected by rate factors.This article mainly introduces traditional claim frequency models and zero-infated models. Through analysising a group of classical insurance data, our resultsshow zero-infated models provide a good ft to the data, the parameter estimationis more accurate, and the conclusion is more tally with the actual situation. Thispaper also ofers a method to solve this kind of problem for the non-life insuranceactuaries in the insurance companies.