摘要
近年来,国内财险公司利用广义线性模型(GLMs)对非寿险业务,尤其是车险业务进行建模和精算分析,使得精算技术人员对保险数据的处理更加细致、科学和公平。基于位置、尺度和形状的广义可加模型(GAMLSS)是GLMs、GAMs、DGLMs和GLMMs等的最新拓展,在介绍该模型的定义、算法和模型实现的基础上,以其框架下的零调整逆高斯模型(ZAIG)为一个特例,讨论了其在财险公司财险定价中的应用研究。最后,以瑞士汽车第三者责任保险的一组损失数据为例进行了实证分析,说明了零调整逆高斯模型在车险费率厘定中是一种较合理的方法,为精算技术人员提供参考和借鉴。
In recent years, domestic P&C companies have used generalized linear models (GLMs) in the modeling and actuarial analysis of non-life insurance business, especially auto insurance business, which makes insurance data processing more meticulous, scientific and fair for the actuarial technical staff. Generalized additive models for loca- tion, scale and shape (GAMLSS) is the latest expansion of the GLMs, GAMs, DGLMs and GLMMs. On the basis of introducing the definition of the models, algorithms and implementations, using the zero adjusted inverse Gaussian regression model (ZAIG) under the GAMI_SS framework as a special case, the paper discussed its application in the study of non-life insurance pricing. Finally,this study empirically examined a classic Swedish data set of third-party automobile insurance claims,and arrived at the conclusion that the zero adjustment inverse Gaussian model was a rational method in automobile ratemaking which could provide reference for actuarial technicians.
出处
《保险研究》
CSSCI
北大核心
2013年第4期43-51,共9页
Insurance Studies
基金
中央高校基本科研业务费专项资金"金融工程与精算学中的定量风险管理统计模型与方法"(NKZXTD1101)
国家自然科学基金面上项目(71271121)的资助
关键词
汽车保险定价
基于位置、尺度和形状的广义可加模型
零调整逆高斯模型
auto insurance pricing
generalized additive models for location, scale and shape
zero adjusted inverse gaussian model