What is this model's recall on the positive class?
Test set: 990 negatives, 10 positives Model: always predicts "negative" Accuracy: 99%
- 0.0
- 0.99
- 0.01
- 1.0
It never predicts positive, so it finds 0 of the 10 positives; high accuracy hides this on imbalanced data.
Play AI & ML ▸