摘要
在一个参数的可选先验类中选择一个合理的先验问题,类似于从参数空间中估计一个恰当参数的问题.基于这一观点,利用贝叶斯分析的后验分布理论,先得出先验的后验分布计算方法,再根据先验的后验分布确定出合理的先验,从而建立了一个基于先验的贝叶斯先验选择方法,它是ML-Ⅱ先验的一个拓广.
Selecting a reasonable prior, distribution problem in a series of selectable prior distributions of a parameter is similar to estimating a suitable parameter in parameter space. Based on this, this paper uses posterior distribution theory of Bayesian analysis to obtain posterior distribution calculation mehtod of prior and then to obtain reasonable prior according to posterior distribution of prior in ordr to establish a Bayesian Prior selection method based on prior which is extended from ML - Ⅱ prior.
出处
《重庆工商大学学报（自然科学版）》
2006年第6期548-550,共3页
Journal of Chongqing Technology and Business University:Natural Science Edition
关键词
先验选择
后验分布
后验似然合理先验
prior selection
posterior distribution
posterior like reasonable prior
