I’ve used the binary scorer and the scorer node to get performance metrics for my random forest classification models. I noticed that the Cohen’s Kappa value from the scorer node is very similar or the same as the MCC value from the binary scorer node. Is this normal?
I’m afraid its a while ago that I did this at uni, so I can’t go into detail on when this is the case, but simply looking at their definitions on Wikipedia suggests to me that they may indeed be similar based on your classification: