有限总体中总体数量的贝叶斯预测
BAYES PREDICTION OF POPULATION QUANTITIES IN A FINITE POPULATION

本文在误差平方损失下研究了具有正态逆伽马先验信息有限总体中的预测问题.首先，基于贝叶斯思想分别获得了线性数量和二次型数量的贝叶斯预测；其次，利用统计决策理论对线性数量的贝叶斯预测和最佳线性无偏预测进行了比较.结果表明在预测均方误差下线性数量的贝叶斯预测一致优于最佳线性无偏预测.
In this paper, we investigate the prediction in a finite population with the normal inverse-Gamma prior under the squared error loss. First, we obtain Bayes prediction of linear quantities and quadratic quantities based on Bayesian theory, respectively. Second, we compare Bayes prediction with the best linear unbiased prediction of linear quantities according to statistical decision theory, which shows that Bayes prediction is better than the best linear unbiased prediction