贝叶斯方法及WinBUGS在非寿险费率分析中的应用  

Application of Bayesian Approach and WinBUGS in Non_Life Insurance Premium Rate

作　　者：吕定海[1] 王晔[2] 

机构地区：[1]南开大学经济学院,天津300071 [2]华农财产保险股份有限公司,北京100034

出　　处：《保险职业学院学报》2013年第1期77-82,共6页Journal of Insurance Professional College

摘　　要：本文主要分析了贝叶斯视角下的广义线性模型,并通过WinBUGS建立分析损失频率的广义泊松模型。考虑到WinBUGS对数据初处理功能的不足,使用R2WinBUGS包交互使用R和WinBUGS。通过一个数值例子,对比说明贝叶斯方法与广义线性模型结果的区别,并指出实际应用中可能遇到的困惑。In this paper, the authors mainly analyze generalized linear models from the perspective of Bayesian method and construct a generalized Poisson model for loss frequency data using WinBUGS. Considering the shortcoming of early data processing for WinBUGS, the authors use R package R2WinBUGS to interactively use R and WinBUGS. Through a numeric example, the authors point out the difference between Bayesian meth- ods and generalized linear models, and the confusion that may be encountered in actual application.

关 键 词：非寿险定价 贝叶斯方法 WINBUGS MCMC 

分 类 号：O212.1[理学—概率论与数理统计] F840.62[理学—数学]