For some reason the command "Save Inverse Prediction Formula", does not save the formula to the associated table. The command "Custom Inverse Prediction" works and does calculate one of the two possible X variable solutions (the function is a Gaussian curve and therefor has two X solutions for each Y variable). But I can not seem to get the formula to save as a new column.
I have read the docs on page 506 in the JMP Stat Graph Guide, and there does not appear to be any steps required other than selecting the command under the red triangle.
I have fit a nonlinear model (Gaussian Peak Shape 3D) to a series of monthly temperature data by site (there are 97 sites, therefore 97 different fits in the graphic output). I performed a "Custom Inverse Prediction" (where a Y value of 5°C = ? X value). From here I have two questions:
1. The "Custom Inverse Prediction" command worked in that it saved one of the two solutions (since the function is a bell shaped curve, there are two X solutions for each Y value). The "Custom Inverse Prediction" command reported only the solution on the right hand side of the bell cure, and not the solution on the left hand side. How can I get both solutions?
2. Trying to "Save Formulas" -> "Save Inverse Prediction Formula" does not seem to work on a data set that has be split "by" some variable. If I perform the same analysis on the full data set (all sites at once). The "Save Inverse Prediction Formula" command works as expected, writing the new columns to the data table. Does anyone know how to get the the formula saved for each site when the analysis is done "by site"?