I'm trying to explore a continuous target variable in SPSS Modeler v. 18.2, using a split variable ("Cohort"). In other models that have a nominal target variable, I'm able to use the auto-classifier to generate models on each split---but in this model when I use the auto-numeric node it ignores the splits entirely. Here is the stream:
In the data file, I have "Cohort" set to Split:
In the node, in the Fields tab, I have added Cohort to the splits...
...and in the Model tab I have checked the build model for each split box:
The nugget doesn't include the splits---in the Summary tab it doesn't look like it's in the model at all:
My work-around is to use Select nodes for each split but that has disadvantages---thank you in advance for any help/corrections.
I am currently using IBM SPSS Modeler 18.0 but I am seeing the exact same behavior when using one of the demo data sets supplied with Modeler. I would consider this to be a defect and something that would need to be addressed by IBM's development team.
I suggest that you replicate the issue with one of the data sets from the "Demos" folder such as the "car_insurance_claims.sav" and then open a support ticket with the IBM SPSS technical support to have this resolved.