nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

non parametric tolerance intervals

Peter_Hersh

I have not been able to find a straight forward way to calculate tolerance intervals if I have non parametric data am I missing something?

Accepted answers

All comments

scwise

Peter,

Would this work? Go to Distribution and fit a Continuous Fit - Smooth Curve to get into a Non-Parametric fit.

Then a red trangle option was to set a K-Sigma "specs" based on quantile from the distribution.

Then you can get the full non-parametric process capability in detail.

10679_Screen Shot 2015-12-21 at 10.44.06 AM.png 10680_Screen Shot 2015-12-21 at 10.44.18 AM.png

Here is an example using Big Class...

Hope this helps!

Scott

Peter_Hersh

Thanks Scott that is a nice option, but not quite what I was hoping for in this case. The customer wants to be able to be able to select a confidence and a proportion to cover like we can currently done in tolerance intervals only using non-parametric data (this is a feature in Minitab which they are hoping to recreate in JMP).

Jordan_Hiller

How about a bootstrap analysis of the Quantiles output from Distribution? Not exactly the same as a tolerance interval, but it would get you the same type of information without parametric assumptions.

JMP Pro only solution, of course. Maybe you can upsell!

Peter_Hersh

Thanks Jordan that is a good idea to get around the parametric assumptions. Maybe an upsell opportunity in the future right now just trying to keep them happy with the conversion from Minitab.

mia_stephens

JMP only provides parametric tolerance intervals, but as Jordan points out, bootstrapping is a nice option. Peter, I have an Excel file that does these computations if it would help - this was from my Six Sigma days. Happy to share.

Peter_Hersh

That would be great. Thanks