試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 64 / 65
Which calibration method, being nonlinear and stable, should be used when a classification model's output probabilities deviate from the actual occurrence probabilities?
①Bagging (bagging)
②dropout
③Platt scaling
④Early stopping
⑤Isotonic regression (isotonic regression)