Capability study [问题]
Often , we need determine our process capability, ( Cp, Cpk, Pp, Ppk. etc. )
( I will take a continuous data for example later . )
The following we need to check first :
1, all data should be under statiticsal control .
ATTENTION: although all your data is within you spec, that still does not means your data is under control .
2, all your data should be normal . It is no meaning to transfer your non-normal data with cox-box tool .
if your data does not meet the above , better not carry out your cpapbility study ! Or your conclusion will leads you to wrong evaluation for your process.
What you should do is :
1, stable your prcess first .
2, Okay, I will discuss more ... tomorrow.
Thanks !
( I will take a continuous data for example later . )
The following we need to check first :
1, all data should be under statiticsal control .
ATTENTION: although all your data is within you spec, that still does not means your data is under control .
2, all your data should be normal . It is no meaning to transfer your non-normal data with cox-box tool .
if your data does not meet the above , better not carry out your cpapbility study ! Or your conclusion will leads you to wrong evaluation for your process.
What you should do is :
1, stable your prcess first .
2, Okay, I will discuss more ... tomorrow.
Thanks !
没有找到相关结果
已邀请:
RSS Feed
5 个回复
Brown (威望:0)
赞同来自:
数据量小于50 分组5～7
数据量500～100 分组6～10
数据量100～250 分组7～12
数据量超过250 分组10～20
但是mini的分组数目是通过计算来确定的。在相同数据数目时，分组数目会有所不同。当然，我们也不必对此过于在意，因为软件已经帮助我们绘制了。并且这对于最终的计算结果没有什么影响。
另外，计算流程能力的分组方式和软件绘制直方图的自动分组方式是相同的。
欢迎别的网友对此发表个人看法。