topic plotting confidence intervals in JMP "fit curve" in Discussions

plotting confidence intervals in JMP "fit curve"

Greeny — Sun, 24 Mar 2024 00:24:33 GMT

Hi,

Is there a way to overlay the confidence interval (shaded) with the fitted curve when using the "fit curve" feature in JMP (without JSL)?

Thanks.

Re: plotting confidence intervals in JMP "fit curve"

jthi — Sun, 24 Mar 2024 07:00:19 GMT

Doesn't seem like you can add those to Fit Curve platform so you might have to use Fit Y by X platform (or maybe something else).

Re: plotting confidence intervals in JMP "fit curve"

Greeny — Sun, 24 Mar 2024 15:33:39 GMT

Thanks. As I am trying to fit a sigmoidal curve (3/4/5-parameter logistic), my only options in JMP-Pro are "Fit Curve" or "Nonlinear". Neither of these platforms provides the option of including shaded confidence intervals with the fitted curve.

Re: plotting confidence intervals in JMP "fit curve"

Ben_BarrIngh — Mon, 25 Mar 2024 09:47:38 GMT

Hi @Greeny,

If you use the Prediction Profiler in Fit Curve that should have confidence intervals - if you're looking for it in the sections of the Fit Curve platform that is not currently available, might be a good point to add to the wish list.

Thanks,

Ben

Re: plotting confidence intervals in JMP "fit curve"

Greeny — Thu, 28 Mar 2024 20:37:49 GMT

Thanks, Ben. Yes, I need the confidence or prediction interval around the fitted curve that appears with the data points overlaid. I don't bother with wish lists in JMP because based on my experience, it hasn't been fruitful. For example, even a simple matter of connecting the medians between boxplots in the Fit Y by X platform which I requested several times many years ago has still not been implemented!

Re: plotting confidence intervals in JMP "fit curve"

Mark_Bailey — Fri, 29 Mar 2024 13:46:55 GMT

Please see this page in the documentation about Nonlinear Platform Options (red triangle). You can save the confidence interval formula and plot it along with the prediction formula in Profiler (Graph menu).