两批产品实际技术参数误差方差之比的多重贝叶斯检验  

Multiple Bayesian Testing Rule on Error Variance Ratio of Practical Technical Parameter in Connection with Two Lots of Products

作　　者：张新育[1] 周世国[1] 

机构地区：[1]郑州大学数学系,河南郑州450001

出　　处：《信阳师范学院学报（自然科学版）》2010年第2期187-189,共3页Journal of Xinyang Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(10671183)

摘　　要：在产品实际技术参数与设计值的差服从位置参数为零的正态分布情况下,对两批分别来自两个总体的误差样本,导出了误差标准差之比的多重贝叶斯检验规则.两标准差参数的联合先验分布采取了边沿Jeff-reys先验.推导的理论依据是后验概率最大原则.文末给出了一个应用实例.Suppose that the difference between the practical technical parameter and the designed standard is normally distributed with zero valued positional parameter, a multiple Bayesian testing rule on the ratio of error standard deviation in accordance with two lots of error sample is presented. The joint prior distribution of two standard deviation parameters is due to Jeffreys＇ , and the theoretical foundation of deduction is the maximum posterior probability rule. Finally, a practical example is given to illustrate how to use this testing rule.

关 键 词：两批产品实际技术参数 误差标准差之比 多重假设 贝叶斯检验 

分 类 号：O212.1[理学—概率论与数理统计]