Akaike's Information Criterion (AIC) not passed from HP GLM to Model Comparison node

[Mike90]

Akaike's Information Criterion (AIC) is shown in the HP GLM model's results (*).  It is not passed on for Model Comparison, as shown in the attachment.  Other tests, such as Average Squared Error, are being passed on.

AIC is set as the model selection criteria.  You can see that AutoNeural is the only model which passes AIC to the Model Comparison node.  As you can see, none of the following models are passing on AIC for model selection: Decision Tree, Random Forest, Gradient Boosting, and HP GLM.  I'm trying to get each model to determine and pass on the AIC test results.  (The tree-based models are not presently showing this test in the individual model results.  That is the subject of a separate post.)

 (*)  As seen in the diagram, AIC is calculated in the GLM model, but it is not assigned to a variable, while Average Squared error is.

[MikeStockstill]

Please contact technical support about this issue.  We recommend this link:

https://support.sas.com/en/technical-support.html#contact-technical-support 

On that page. select Create A Track, and follow the steps.

Thank you.