topic Grouped Contour Plot of Lower 95% Prediction Interval in Discussions

Grouped Contour Plot of Lower 95% Prediction Interval

JoffJones — Fri, 16 Jun 2023 12:25:08 GMT

Hi, I am trying to plot contour plots for the lower 95% prediction interval that are grouped by two variables. I can achieve the plot in Graph Builder, but it does not display the predictions for design points that have no experimental data. What I would like to do is plot all of the predictions over the entire possible range of my variables.

Is there a way to generate the predictions for a grid of values for your independent values? Is this achievable using the GUI?

Many thanks,

Joff

Re: Grouped Contour Plot of Lower 95% Prediction Interval

MRB3855 — Fri, 16 Jun 2023 13:32:54 GMT

Hi @JoffJones One way to do this is to include extra rows in your data table with the x (however many x’s you have) values you want intervals for…but for those rows leave the response as missing. Then, fit your model and save the interval formula to your table. That will result in intervals at all values of x you specified (the design points and the additional x vales).

Re: Grouped Contour Plot of Lower 95% Prediction Interval

JoffJones — Fri, 16 Jun 2023 13:48:32 GMT

Thanks @MRB3855 ! Is there any easy way to add all the additional x values? I guess that I could generate them by creating a new full factorial design and pasting them in - but maybe there is a better way.

Re: Grouped Contour Plot of Lower 95% Prediction Interval

MRB3855 — Sat, 17 Jun 2023 13:02:07 GMT

@JoffJones for the purposes of filling the entire design space so the contours look smooth, I'd choose a Space Filling Design (Uniform Method) rather than a factorial.