Applying models to new data sets

[#1]

Hi all,

I have creating a model from a particular DOE and wish to apply that model to a new data set from another experiment (parameters are all teh same, but factors may be different).  Is this possible?

[#2]

You can save any model fitted by JMP platform as a new column with a formula. Click the red triangle next to the platform or model and select Save or Save Columns to see your choices. You can add rows with new factor levels or copy the formula to another data table.

[#3]

Hi markbailey,

I can save the prediction formula but I want to use the prediction profiler with the DOE factors and model but the response from the new experiment. However, the factors of the new experiment don't exactly align to that of the DOE. To be more specific, the DOE examined the response of a device to all of the important factors. The new experiment examined the response of the same device to the exclusion of a processing step. I want to use the DOE model with a prediction profiler for this new data... If that makes sense.

[#4]

You have selected a model for the response in one experiment. The results in the Pediction Profiler are now based on the model, not the data. That is, the behavior of the profiles won’t change unless the model changes. Using the model with new data (I.e., the second experiment) will produce the same profiles.

What do you mean by, “the factors of the new experiment don’t exactly align to that of the DOE?”

If you mean by “to the exclusion of the processing step” that some of the original DOE factors were not varied but held at a fixed level? You can simply add data columns for those fixed factors and populate them with the given level. You could also replace the exluded factor column in the formula with the fixed level.