medium10 pts · first answer countsmodel-evaluationclassification-metricsimbalanced-data

Weighted F1 on imbalanced classes

A report shows weighted F1 = 0.88 for a 3-class model. Why can this hide poor performance on a small class?

A.Weighted F1 leaves out the largest class entirelyB.It only uses precision and ignores recallC.Class F1s are weighted by support, so big classes dominateD.It is always computed on the training data

Discussion 0

Join free to post in the discussion.

Weighted F1 on imbalanced classes

Discussion 0

Related questions