不平衡单因素随机效应模型下过程无能力指数的区间估计  

Interval Estimation of Process Incapacity Index in Unbalanced One-Way Random Effect Model

作　　者：贺加贝 李新民 HE Jiabei;LI Xinmin(School of Mathematics and Statistics,Qingdao University,Qingdao 266071)

机构地区：[1]青岛大学数学与统计学院,青岛266071

出　　处：《系统科学与数学》2020年第2期281-288,共8页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金(11871294)资助课题。

摘　　要：过程能力分析在工业生产上是衡量生产能力以及保障产品质量的一个重要问题.与其他指标相比,过程无能力指数C(pp)不仅能直接反映过程的不可靠性,而且能提供更多关于过程不精确性的相关信息.因此,它可以帮助供应商改进不良过程来提高质量性能.在采样的数据通常是来自于不平衡单因素随机效应模型时,文章利用近似广义推断、Fiducial推断和参数自助法分别给出了C(pp)的区间估计,并对3种区间估计的性能进行了模拟比较.模拟结果表明Fiducial区间估计优于近似广义推断和参数自助法,特别是在数据不平衡时.Process capability analysis is an important problem in industrial production to measure production capacity and guarantee product quality.The Cpp not only can directly reflect the process incapability,but also provides more information regarding process inaccuracy than other indices.Therefore,it can help contract manufacturers to better understand their process for improving poor quality performance.When the sampling data are collected from unbalanced one-way random effect model,approximate generalized confidence interval,fiducial generalized confidence interval,parametric bootstrap confidence interval of Cpp are proposed respectively.In addition,simulation studies are conducted to compare the performance of interval estimations.Simulation results show that the fiducial generalized interval is better than other intervals,especially for very unbalanced designs.

关 键 词：过程无能力指数 广义推断 Fiducial推断 参数自助 不平衡数据 

分 类 号：O212.1[理学—概率论与数理统计]