I want to "stretch" the confusion matrix of a binary classification back to some categorical variable in the initail data. Is that possible?
What I have: Pr