Evaluating established models in KNIME



  I have established quite a few predictive models in KNIME and I was looking forward to test the quality of these models using the "Akaike information criterion". I was wondering if there is any node in KNIME that gives me the ability to run this measure in order to compare my models. If not, how do you compare what number of columns you should include in your model?




You can try the Backwared Feature Elimination metanode.

Yes it is working alright! Thank you Thor.