基于区间数回归模型的未决赔款准备金评估  

Evaluating the outstanding claim reserve with interval regression

作　　者：冯卫泽 王达布希拉图 

机构地区：[1]广州大学数学与信息科学学院,广东广州510006 [2]广州大学经济与统计学院,广东广州510006 [3]广州大学岭南统计研究中心,广东广州510006

出　　处：《广州大学学报（自然科学版）》2015年第1期12-17,共6页Journal of Guangzhou University:Natural Science Edition

基　　金：国家自然科学基金资助项目(11271096)

摘　　要：区间数具有信息含量丰富、界限分明的优点,在实数型数据信息不完全的场合用区间数替代可提高回归模型的预测效果.非寿险未决赔款准备金是对保险公司的一项重要负债,其评估方法和效果对保险业发展至关重要.文章在链接法的指数平滑型进展因子基础上,利用区间数回归模型估计各发展年的进展因子,给出一种新的未来未决赔款准备金的估计方法.Interval data have the advantages of rich information and clear boundaries, in the case of real data with incomplete information, interval regression has better prediction effect than traditional regression. Out- standing claim reserve is an important liability of insurance companies, the evaluation method and the effect of outstanding claim reserve is crucial to the development of insurance business. In this paper, based on the chain ladder method, interval linear regression model is used to estimate the development factor, which gives a new method for estimating outstanding claim reserve.

关 键 词：准备金 区间数回归 链梯法 

分 类 号：F840.48[经济管理—保险]