主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2004, Vol. ›› Issue (5): 30-34.

• 论文 • 上一篇    下一篇

二维GPSJ分布类及其在保险中的应用

高洪忠1, 任燕燕2   

  1. 1. 中国人民财产保险股份有限公司博士后工作站 北京 100052;
    2. 山东大学经济学院 济南 250100
  • 收稿日期:2004-03-23 修回日期:2004-08-31 出版日期:2004-10-28 发布日期:2012-03-07

The Bivariate GPSJ Class and Its Application in Insurance

GAO Hong-zhong1, REN Yan-yan 2   

  1. 1. PICC Property and Casualty Company Limited, Beijing 100052, China;
    2. Institute of Economics, Shandong University, Jinan 250100, China
  • Received:2004-03-23 Revised:2004-08-31 Online:2004-10-28 Published:2012-03-07

摘要: 本文给出的二维GPSJ分布类(即GPSJ2)在保险领域应用范围较广,如由机动车辆保险由财产损失索赔次数和人身伤亡索赔次数所组成的模型、超额再保险中的由自留保单数和分出报单数所组成的二维模型等。本文在GPSJ1类基础上研究二维GPSJ类,本文给出了此分布类的定义、背景意义、概率计算的递推公式,同时还对各参数的最大似然估计值的求解过程进行了分析。本文最后将这一模型应用于一组机动车辆赔付次数数据,考虑了由财产损失索赔次数、人身伤亡索赔次数组成的二维模型,拟合效果令人满意。

关键词: GPSJ类, EDP变换, ESJ类, GPSJ2分布类

Abstract: The model presented in this paper is very useful in insurance,especially for frequencies of property claims/frequencies of liability claims model,frequencies of retention claims/frequencies of excess treaty claims model in reinsurance.We study a class of bivariate GPSJ distributions by extending the GPSJ class from the uni-variate case to the bivariate case.We also show how to obtain the recursive formula of its probabilities.The study typically extends the use of some bivariate distributions appeared in literatures.Finally we employ this class to an application and the fitting effect is satisfying.

Key words: GPSJ1class, extended De Pril transform, ESJ class, GPSJ2class

中图分类号: