摘要
控制图的报警阀值是由异常点来启动的,而异常点的判断则取决于小概率事件的发生,本文给出了用样本均值和方差去估计总体均值和方差时,点落在控制界限之外的分布函数。使用本文所给的分布函数去计算小概率事件避免了用样本特性值代替总体特性值时所带来的误差,同时也统一了控制图在不同异常点出现时的报警阀值。
The alarming boundary values of control chart is motiveted by the extraordinary point ,however,the determination of the extraordinary point depend on the event with small probability. In this paper,we give the distribution density for the point falling outside of upper and lower bound when we use average and standard deviation in sample to estimate the eversge and standard deviation in population. Using this distribution density to calculate event with small probability,we can aviod the error that is caussed by using sample characteristic for calculating but using totality characteristic for calculating. we unify the boundary values of alarming in this paper.
出处
《湖北大学学报(自然科学版)》
CAS
1994年第4期396-404,共9页
Journal of Hubei University:Natural Science
关键词
小概率事件
报警阀值
控制图
样本特性值
Event with small probability Boundary values of alarming Upper and lowerbound