topic Re: Export confidence intervals of th profiler in Discussions

Export confidence intervals of th profiler

Julianveda — Thu, 06 Jul 2023 14:43:54 GMT

Hi Community,

I was wondering if it is posible to export the information from the profiler, mainly the confidence intervales (gray region in image below). It seems to me there is a way to do it but with the Pro version. Do you know if there is a way to do it with the JMP basic version?

Re: Export confidence intervals of th profiler

Victor_G — Thu, 06 Jul 2023 16:06:36 GMT

Hi @Julianveda,

Depending on the platform you're using to build the model, it can be easily accessible.
Assuming the profiler comes from the "Fit Model" platform, you can very easily (in the red triangle next to your response name) go to "Save Columns" and then "Mean Confidence Limits Formula" (and "Prediction Formula"):

If you want to re-create the Profiler with the formulas, you need to have the "Prediction Formula" and the "StdErr Pred Formula" saved in your datatable. Then, going into "Graph", "Profiler", you can use these columns to recreate your profiler (click on Yes when a window pops up asking you to use PredSE to create intervals for your response) :

I hope this answer will help you,

Re: Export confidence intervals of th profiler

Julianveda — Mon, 10 Jul 2023 15:30:41 GMT

Hi @Victor_G

Thank you very much . This is not exactly the answer to my question, but I accepted it as a solution because your explantations helped me a lot anyway.

Just in case you want some clarification about my initial objective: In a certain way, what I wanted was to reproduce exactly the profiler in Excel for other people not having JMP. With the predicted formula you can easily create the fitting curve, but I also wanted to include the gray regions that we see in the profiler.

With your explanation I could present things a little bit different, but it also does the job: I plotted actual vs predicted and use the columns with the formula for confidence intervals. (this is slightly different from my objective which is use the original graph and put the fitting curve and confidence interval as the profiler does).

Thank you again, because your solution provides the majority of elements to answer my question.