Re: Cauchy distribution. Where is Process capability button?
This widget could not be displayed.
Created:
Aug 12, 2020 08:58 AM
| Last Modified: Aug 12, 2020 5:58 AM(2318 views)
| Posted in reply to message from Leonard 08-12-2020
According to the JMP Documentation, the Cauchy distribution has an undefined Mean and Standard Deviation. Based upon that, I assume JMP does not have Cp and Cpk for that distribution and thus no need for the button.
Re: Cauchy distribution. Where is Process capability button?
This widget could not be displayed.
Created:
Aug 12, 2020 08:58 AM
| Last Modified: Aug 12, 2020 5:58 AM(2319 views)
| Posted in reply to message from Leonard 08-12-2020
According to the JMP Documentation, the Cauchy distribution has an undefined Mean and Standard Deviation. Based upon that, I assume JMP does not have Cp and Cpk for that distribution and thus no need for the button.
Re: Cauchy distribution. Where is Process capability button?
This widget could not be displayed.
Created:
Aug 12, 2020 09:33 AM
| Last Modified: Aug 12, 2020 6:33 AM(2315 views)
| Posted in reply to message from txnelson 08-12-2020
Yes, but then how do I calculate PPK? Should I take the nearest distribution, if it is? Or should I calculate PPK using Standard Deviation? I need to evaluate the process with nonnormal data distribution.
Re: Cauchy distribution. Where is Process capability button?
This widget could not be displayed.
Created:
Aug 12, 2020 09:41 AM
| Last Modified: Aug 12, 2020 6:41 AM(2312 views)
| Posted in reply to message from Leonard 08-12-2020
Have you tried fitting all the Distributions, and then determine which have a significant Goodness Of Fit? You could then use that distribution for your Cp and Cpk
Re: Cauchy distribution. Where is Process capability button?
This widget could not be displayed.
Created:
Aug 12, 2020 02:01 PM
| Last Modified: Aug 12, 2020 11:01 AM(2262 views)
| Posted in reply to message from txnelson 08-12-2020
Actually, the mean of the Cauchy distribution is defined but not the higher moments. So, no variance, no standard deviation. No standard deviation, no Ppk.
And, just as an aside, if you're responsible for improving a process that outputs data that are really Cauchy distributed, you might want to update your resume...