Add a processing layer for ADWINOutcome
tms-bananaquit opened this issue · 0 comments
tms-bananaquit commented
Rather than always checking y_true == y_pred
for accuracy, a user might want to have some more elaborate function of those inputs.
This can already be accomplished via change_detection.ADWIN
, but it'd make it look really wonky to pass a running recall_score
calculation via the X
argument or something.