問題文
The pipeline computes the mean and the standard deviation of billed amount, uses them to rescale the column, and then splits the data into fitting and evaluation portions. A reviewer objects to the order. Which claim about the two stages is valid?
選択肢
- The two statistics must be computed on the training portion, and applying them to the other portion is what keeps the enumeration guarantee that evaluation depends on from being contaminated.
- The two statistics must be computed on the pooled portion, and applying them to the other portion is what keeps the enumeration guarantee that evaluation depends on from being contaminated.
- The two statistics must be computed on the training portion, and applying them to the other portion is what keeps the separation guarantee that evaluation depends on from being contaminated.
- The two statistics must be computed on the pooled portion, and applying them to the other portion is what keeps the separation guarantee that evaluation depends on from being contaminated.