JMP Wish List
We want to hear your ideas for improving JMP. Share them here.I did functional data DOE with current JMP Functional data explorer platform. I can get very good model with Functional DOE. in red triangle beside Functional DOE analysis, we have save columns and with that, we have the options available to save prediction formula and save residual formula.
This is very good because I can use the prediction formula to create prediction profiler in Graph > Profiler.
However, when I try to save the std error of predicted formula, I can only go to Generalized regression for FPC scores tab and choose Normal Best subset with AICc validation and find save column and then choose std error of predicted formula. But this function does not give me the std error of predicted formula in my data.
Please add the ability, under save columns, under the Functional DOE analysis tab to save the PredSE formula for the functional DOE prediction (in addition to the save prediction formula option), because this will help me to show the correct prediction function in the Prediction Profiler.
This offers two unique advantages.
1. With the predSE formula, the confidence intervals around the prediction can then be constructed in Graph > Profiler, with the Response as a function of the Predictors. This is a critically important expression of the Functional DOE problem statement (not just the FPCs as a function of the Predictors inside the GenReg for FPC Scores Outline) because it provides directly interpretable estimates of uncertainty around the model Predicted Response (as a function of the Predictors).
2. Functional DOE analysis does not always show the Prediction Profiler even inside the FDE Platform, especially when we have constraints with our design factors or design space is limited. Building the Profiler in Graph > Profiler offers a nice alternative with the saved Prediction Formula (and Saved PredSE Formula) .