首页|期刊导航|井冈山大学学报(自然科学版)|共轭先验下广义指数分布形状参数的Bayes分析

共轭先验下广义指数分布形状参数的Bayes分析OA

Bayes analysis for the shape parameter of generalized exponential distribution under conjugate prior

中文摘要英文摘要

基于加权p,q 对称熵损失函数,研究了广义指数分布形状参数的 Bayes 估计问题.本研究首先在 Gamma先验分布下,给出了形状参数 Bayes 估计的一般形式与精确形式,并证明了该估计具有唯一性和可容许性.其次,进一步推导了形状参数的多层Bayes 估计、经验 Bayes 估计及刀切 Bayes 估计的精确形式.最后,应用 R 软件和MCMC算法进行数值模拟,结果表明:共轭先验下的刀切 Bayes 估计优于传统 Bayes 估计,且稳健性较好;与若干其他损失函数相比,该损失函数下所得 Bayes 估计的精度更高.

The Bayesian estimations problem of the shape parameter in generalized exponential distribution under the weighted p and q symmetric entropy loss functions were explored.The general and precise forms of Bayesian estimation of the shape parameter with the conjugate gamma prior distribution was obtained,and the uniqueness and admissibility of the estimation were proved.On this basis,the exact forms of the multi-layer Bayesian estimation,empirical Bayesian estimation,and jackknife Bayesian estimation of the shape parameter were further derived.Finally,numerical simulations were conducted using R software and MCMC algorithm.The results show that the jackknife Bayesian estimation under the conjugate prior is superior to the traditional Bayesian estimation and has better robustness;compared with several other loss functions,the Bayesian estimation derived under this specific loss function exhibits higher accuracy.

孙霈榆;徐宝

吉林师范大学数学与计算机学院,吉林,四平 136000吉林师范大学数学与计算机学院,吉林,四平 136000

数理科学

广义指数分布加权p,q对称熵损失函数共轭先验分布Bayes估计MCMC算法

generalized exponential distributionweighted p,q symmetric entropy loss functionconjugate prior distributionBayesian estimationMCMC algorithm

《井冈山大学学报(自然科学版)》 2026 (5)

1-11,11

国家自然科学基金项目(11571138)吉林省科技发展计划项目(YDZJ202201ZYTS622)

10.3969/j.issn.1674-8085.2026.05.001

评论