問題文
The first report says the model is correct on ninety-nine and six tenths of claims. A reviewer points out that answering never stalls on every claim would score the same. What should the practitioner conclude?
選択肢
- Quoting the plain share of correct calls is misleading here because of the severe imbalance, so the report must instead carry the correctness value that documentation pairs with recall.
- Quoting the plain share of correct calls is misleading here because of the mild imbalance, so the report must instead carry the positive value that documentation pairs with recall.
- Quoting the plain share of correct calls is misleading here because of the severe imbalance, so the report must instead carry the positive value that documentation pairs with recall.
- Quoting the plain share of correct calls is misleading here because of the mild imbalance, so the report must instead carry the correctness value that documentation pairs with recall.