bug in decision tree

Hi,

When you use the Knime's decision tree, you can obtain fractional number of subjects in leaves of the tree as in the joined picture. I remember I used this technique many times under SPPS that was never the case. Where does it come from ?

Best regards

Fabien

It may be possible that the counts are being generated from the % rather than being tracked independently?

I don't get how a fraction could be valid here. You could try using the selection functionality in an interactive table to see the datapoints present to check the real value. To do this in the table select Hilite -> Filter -> Show hilited only and in the tree select a node and highlight it.

 

 

Hi, thank you for your reply. I thought too that counts where generated from the %, but I really don't see the point. I noticed this bug for the first time a long time ago I really can't figure out why it hasn't been corrected. There must be some reasons. I'll try your solution waiting.

Best regards

Fabien