In discriminant analysis, what is the name of the data set that is used for making predictions?

• A. scoring data set
• B. training data set
• C. validation data set
• D. test data set

Answer: (A)

Scoring data set is what you use to make predictions with a trained discriminant model. You build the discriminant function from the training data, and after you’ve chosen and validated the model, you apply it to new observations to assign class labels or probabilities. This act of applying the model to data to generate predictions is called scoring, so the data set used for that purpose is the scoring data set. The other sets have different roles: training data is used to fit the model, validation data helps tune and compare models, and test data provides an unbiased estimate of performance on unseen data.