贝叶斯定理在合并样本方差估计中的应用  

The Application of Bayesian Theorem to Pooled Estimation of Variance

作　　者：宋明顺[1] 王伟[1] 

机构地区：[1]中国计量学院管理学院,浙江杭州310018

出　　处：《计量学报》2008年第2期186-189,共4页Acta Metrologica Sinica

基　　金：国家自然科学基金(50575215)

摘　　要：合并估计是计量学中的一种常见情况,《测量不确定度表示指南》(简称GUM)推荐了在此情况下估计测量不确定度的常规方法,但它的不足之处在于没有运用先前的测量信息。另一种被称为频率方法则充分利用了先前的测量信息,但在实际中却难以操作。而根据贝叶斯定理推导出的贝叶斯方法克服了上述方法的缺陷和不足,可以给出更可靠的测量结果,并且具有较好的可操作性。The pooled estimating situation is a common case in metrology. The method of estimating measurement uncertainty applied in this case is recommended in the Guide to the Expression of Uncertainty in Measurement（GUM）, which is named as the general method. The general method has a defect that the prior measurement information is not adopted. Another method , named as the frequency method, uses the prior measurement information sufficiently, but it is impossible to operate in practice. The Bayesian method is derived from the Bayesian theorem, which overcomes the shortcomings and defects caused by the methods aforementioned, gives more reliable measurement result, and it is easy and simple to operate in practice.

关 键 词：计量学 贝叶斯定理 合并估计 测量不确定度 

分 类 号：TB9[一般工业技术—计量学]