To get an "honest" or true estimate of the classification error in a binary prediction model, the classification matrix should be constructed using:

Group of answer choices

The training data

The validation data

All of the data