試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 43 / 65
When building a churn prediction model at an insurance company, which method is the most appropriate way to address class imbalance in the training data (few churners)?
①Combine upsampling of minority classes, downsampling of majority classes, and weighting of the loss function.
②Optimize solely using accuracy as the evaluation metric.
③Remove the minority classes and train using only the majority.
④Use the pre-trained model as is.
⑤Use only publicly available data as a substitute.