Mathematical Sciences: The Application of the Stepwise BayesTechnique to Some Statistical Questions

在有限总体抽样情况下，我们将研究估计量的可容许性和它的贝叶斯度对某些（可能不适当的）先验分布的密切关系。经典程序的贝叶斯解释，如极大似然估计，将提供来自有限种群的样本。逐步贝叶斯技术将用于为有限总体抽样程序提供非信息贝叶斯证明，这反映了统计学家的先验信念，即人口中靠近的成员比远离的成员更相似。

The close relationship between the admissibility of an estimator and its Bayesness with respect to some (possibly improper) prior distribution will be studied in the finite population sampling situation. A Bayesian interpretation of classical procedures such as the maximum likelihood estimator will be provided for samples from finite populations. The stepwise Bayes technique will be used to provide a noninformative Bayesian justification for procedures in finite population sampling which reflect the statistician's prior beliefs that members of the population which are close together are more alike than those which are far apart.