金币
UID187217
帖子
主题
积分64
注册时间2014-12-29
最后登录1970-1-1
听众
性别保密
|
欢迎您注册蒲公英
您需要 登录 才可以下载或查看,没有帐号?立即注册
x
pda 13# 报告 4.4警戒和行动限 方法C
c. Nonparametric Tolerance Limit Approach
As environmental monitoring data, especially in cleanrooms, is typically not normally distributed (i.e., exhibits skews heavily toward lower counts or zero counts), a nonparametric tolerance limit approach to setting alert and action levels is recommended. These limits allow the user to assert with at least 95% (K=0.95) confidence that 100(P) or 99% of a population lies below the value, as determined by the stated action limits, for the respective data (16). For distribution-free tolerance limits, minimum sample sizes are N=60 for 95/95 (alert limit) and N=300 for 95/99 (action limits).
没有仔细翻译,大概的看了下:说到测试值,比如微生物什么的大多为小数字或趋向于零,可以考虑95%和99%的分位数方法。
问题是如果小数字或者0够多,得到的警戒和纠偏限就会足够小。C级区测试得到的数据统计并计算后得到警戒和纠偏限为1和3,那不是稍有检出就会报警或超限吗?
请大家积极发烟冒泡~~~
|
|