Let's take this well known example. How could I get the Kappa score and the Matthews correlation coefficient from nested cross-validation?
I've tried to make it with cross_val_predict
instead of cross_val_score
, but I found out that the results of the two are not the same, and since I already have results with cross_val_score
, I think I'd prefer to continue using it.
These are the libraries I'm using:
from sklearn.metrics import matthews_corrcoef
from sklearn.metrics import cohen_kappa_score