摘要
在一个参数的可选先验分布类中选择一个合理先验的问题,类似于从参数空间中估计一个恰当参数的问题。因此,可利用贝叶斯分析的后验分布理论,先求出参数的后验分布,再根据后验分布中各个先验的相对似然选取似然最大的先验为合理先验,从而建立一个基于参数的后验分布的先验选择方法,它也是ML-Ⅱ先验的一个拓广。
Selecting a reasonable prior density in a series of selectable prior distributions of a parameter is similar to estimating a suitable parameter from parameter space. From this point, posterior distribution of Bayesian analysis can be used to solve posterior distribution of a parameter, then reasonable prior of maximum likelihood can be selected by depending on the relative likelihood of each prior of posterior distribution, a prior selection method of posterior distribution based on parameters is established. This method also extends ML - Ⅱ prior.
出处
《重庆工商大学学报（西部论坛）》
2007年第5期67-69,共3页
Journal of Chongqing Technology and Business University:West Forum
关键词
先验选择
后验分布
贝叶斯似然合理先验
prior selection
posterior distribution
Bayesian Likelihood Reasonable Prior
