Decision Trees / Forests and Backward Feature Elimination not possible?

In KNIME 3.2.0, the Random Forest and Gradient Tree complain during the 2nd iteration of the Backward Feature Elimination Metanode that columns are missing. This also happens when pattern matching is used instead of manual column selection. The learner should use all columns available and ignore when any of them is missing during exection.

I think BFE is a cool feature and Trees / Forests are very powerful, so I would really love to see the results with them.

Is that issue already known to you?

Hi Windowsfreak,

we published a small minor bugfix release yesterday evening. Could you update your version of the KNIME Analytics Platform and try again?

(File -> Update KNIME...)

Best regards, Iris