DecisionTree: exclude learning columns but still show them...

...in the table of classified data?

With the column filter node it is easy to exclude columns as input to the decision tree learner. But some of the columns like date would be very useful for result analysis to have later in the table of classified data.

Is there any way to do so?

Thank you.

Frank

Hi,

yes, it is. Just don't exclude those columns before Decision Tree Predictor node. Decision Tree Predictor node will copy input to output and add some additional columns - prediction, probabilities etc.

But that solution will not work for every Predictor nodes. For example, for Weka Predictor it is safer to use Column Splitter node before Predictor and and Column Appender after.