Stein损失函数下的保费估计
Premium Estimator under Stein Loss

利用信度理论研究在Stein损失函数下的保费估计问题,分析了贝叶斯估计、信度估计、多层贝叶斯估计3种估计,并计算3种保费估计,比较其结果发现在Stein损失函数下:当样本数n趋于无穷大时,信度保费会收敛到风险保费,且多层贝叶斯估计比贝叶斯估计的稳健性更强.
The premium estimate under a typical asymmetric Stein loss function is studied by the credibility theory.The three type of estimator including Bayes estimator,credibility estimator and hierarchical Bayes estimator under Stein loss function are discussed.Finally,by numerical simulation method the quality of three estimates are compared.The results show that,under Stein loss function,when the sample size n tends to infinity,all the three premium estimator convergence to risk premiums respectively.In addition,the robustness of hierarchical Bayes estimator is better than that of two other estimator