試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 49 / 65
What is the generally recommended approach for handling features that contain many missing values?
①Remove rows with missing values across the board.
②Always replace in bulk with the average value.
③Investigate the causes and distribution of missing data and use imputation methods or model-based imputation appropriate to the situation.
④Only add a flag indicating missingness and leave the values unchanged.
⑤Reduce dimensionality with principal component analysis before training.