Prediction and Tolerance Interval for the Profiler Tool

[Marco_]

The Profiler is a powerfull tool in JMP and widely used for the analysis and interpretation of Design of Experiment models.

To make prediction about future runs, the uncertainity of a model has to be considered. However, JMP is currently limited to confidence intervals, which only allow to make predictions about the range where the true mean falls into. Prediction and tolerance intervals on the other side would extend this so single runs or even for a poportion of a population.

Other software tools such as DesignExpert offer these options already for years and I guess many other JMP users would appreacitae such an option in one of the future JMP versions.

[Marco_]

@Sarah-Sylvestre Thanks for pointing that out. I was a bit afraid that I have overlooked an older post in the community.

Since you evaluating this feature now. It would be important to offer PIs and TIs not only in the 'Standard Least Squares' platform but also in the Mixed Model and Generalized Regression platform. Moreover, please allow the export of these intervals in form of formula columns as it is currently already possible with CIs.

[Sarah-Sylvestre]

@Marco_ @ih Status is changed to investigating as we are looking into possibly implementing this in a future release.  Changed status from "duplicate" because there is no duplicate wishlist item, as this is a duplicate of an internal tech support track. 

[Sarah-Sylvestre]

@Marco_ Thank you for all of your suggestions related to adding prediction and tolerance intervals throughout JMP! Some of the suggestions you made in the comments are not related to the prediction profiler, so if you haven't already, please submit a separate wishlist item for those requests.  We like to have 1 request per wishlist item if possible. Thank you! 

[Steffen_Bugge]

This is a great suggestion from @Marco_ ! Please implement this asap!

[hogi]

@Sarah-Sylvestre , thanks for delivering with JMP18.

[Sarah-Sylvestre]

Available in JMP 18. 

[plebrun]

Many thanks for having added prediction intervals in the profiler !

Next step would be to use the same distribution in the Simulator ! So far, only a Normal is available, not accounting for the design support, which makes it very optimistic (remember in DoE, degrees of ofreedom are often low). A Student's t with d.f. from the model and the information from the design matrix (i.e. the same as in the prediction interval) could be used ? I raised this request few years ago (along with the PI in the profiler) so maybe is it in the pipe ?

Thanks,