一种对称损失下逆高斯分布形状参数的Bayes估计
Bayes estimation of shape parameter in inverse Gaussian distribution under a symmetric loss
孙双 1徐宝1
作者信息
- 1. 吉林师范大学数学与计算机学院,吉林四平 136000
- 折叠
摘要
在加权p、q对称损失下,分别研究了逆高斯分布形状参数的Bayes估计、多层Bayes估计和E-Bayes估计,并把刀切法的思想运用到Bayes估计中,得到逆高斯分布形状参数的刀切Bayes估计的精确形式,为验证形状参数估计的合理性,运用R软件,采用随机游动Metropolis算法对所研究参数的Bayes估计、E-Bayes估计和刀切Bayes估计进行数值模拟,比较了在加权p、q对称损失、Linex非对称损失、平方损失和q-对称损失下逆高斯分布形状参数的Bayes估计的精度,结果表明加权p、q对称损失下逆高斯分布形状参数的Bayes估计的精度最高.
Abstract
Bayes estimation,multilayer Bayes estimation and E-Bayes estimation of inverse Gaussian distribution parameters are studied respectively under weighted p and q symmetric entropy loss,and the idea of knife cutting method is applied to Bayes estimation.The exact form of Bayes estimation of shape parameter of inverse Gaussian distribution is obtained.R software is used to numerically simulate Bayes estimation,E-Bayes estimation and Jackknife-Bayes estimation of studied parameters by u-sing the random walk Metropolis algorithm.The Bayes estimation precision of the shape parameter of inverse Gaussian distribu-tion under the following loss:weighted p and q symmetric entropy loss,Linex asymmetric loss,square loss and q-symmetric entropy loss is compared,and the results showed that the Bayes estimation of the shape parameter of the inverse Gaussian dis-tribution has higher accuracy under weighted p and q symmetric entropy loss.
关键词
Bayes估计/逆高斯分布/刀切法/损失函数/随机游动Metropolis算法Key words
Bayes estimation/inverse Gaussian distribution/Jackknife/loss function/random walk Metropolis algorithm引用本文复制引用
基金项目
国家自然科学基金资助项目(11571138)
吉林省科技发展计划项目(YDZJ2022ZYTS622)
出版年
2024
NETL
NSTL
万方数据