試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 43 / 65
Which method is most appropriate when you want to quantitatively identify, feature by feature, the causes of a model's biased predictions toward a specific department?
①Try to improve the model by retraining it repeatedly.
②Reassess after collecting additional data for each department.
③Extract misclassified samples and manually verify the labels.
④Visualize feature contributions using explainability methods such as SHAP and LIME.
⑤Move the model to the cloud and scale it.