摘要
由于推导过程能力指数的统计分布通常是困难的,特别是对数据不服从正态分布和过程有偏离情形下过程能力指数的统计分布研究。文章给出了基于Boostrap重抽样技术的过程能力指数置信区间的大数据构建方法,该方法对构造过程能力指数的置信区间有一定自适应性和普适指导性,为正确理解与使用过程能力指数提供评判依据。
It is usually difficult to derive the statistical distribution of process capability index,especially when the data is not normally distributed and the process is deviated. In view of this,the paper presents big data method of constructing confidence interval of process capability index based on Boostrap re-sampling technique. The proposed method has certain adaptability and universality to construct the confidence interval of process capability index,providing a judgment basis for correctly understanding and using process capability index.
引文
[1]Efron B.Bootstrap Methods:Another Look at the Jackknife[J].The Annuals of Statistics,1979,7(1).
[2]Pearn W L,et al.Distributional and Inferential Properties of PCIs[J].Journal of Quality Technology,1992,24(4).
[3]Pearn W L,et al.Making Decisions in Assessing Process Capability Index Cpk[J].Qual.Reliab.Engng.Int,1999,(15).
[4]Pearn W L,Lin P C.Measuring Proeess Yield Based on the Capability index Cpm[J].1nt J Ady Manuf Technol,2004,(24).
[5]生志荣.过程能力分析与过程能力指数有关研究[D].上海:华东师范大学学位论文,2007.
[6]常振海,刘薇.非对称分布的非参数bootstrap法置信区间比较[J].统计与决策,2014,(20)./
(1)为了使读者能重复本文的重抽样实验结果(B表示重抽样次数),这里设定R语言种子数为set.seed(123),需要R语言代码可与作者联系。