Decision Tree Learner does not work

KNIME Analytics Platform
1

Dear all,

I have some string data (Errors) denoted 0 and 1 and some variables with z-score normalized numbers. I try to make explain the Error if it is occuring (=1) with my other variables. If I try to employ the Decision Tree learner it always fails with the warning “WARN Decision Tree Learner 0:420:537 Class column “0” not found or incompatible”. (It is called column 0 as I installed a Backward Feature Elimination Loop around the Decision Tree).
May it be that the Decision Tree Learner Fails because the first Error Column is nearly always 0? If so, why does it not just continues with the next column and tells me, it has no Output for the first column? I really do not know how I can resolve this issue. Thanks in Advance for any help.

Kind regards, Lisa :slight_smile:

2

Hello Lisa,

the Decision Tree Learner tells you that your class column is missing (column 0 in your case).
If you do Feature Elimination, you have to ensure that the class column is among the static columns (i.e. is contained in each loop iteration) as you are looking for the feature set that best explains the class columns.

Kind regards,

Adrian

3

The Column should be static. I did the same Loop with another learner and it worked.