Which statement best describes testing data in machine learning?

Which statement best describes testing data in machine learning?

• A. Testing data is used to determine how well the model is performing.
• B. Testing data includes answers.
• C. Testing data does not include the model's answers.
• D. Testing data is the same as training data.

Answer: (C)

Testing data is the set you use to see how well a trained model generalizes to new, unseen examples. Its defining trait is that the data itself does not contain the model’s predictions; you generate those predictions by running the model on the test inputs and then compare them to the true labels to measure performance. This separation prevents data leakage and gives a honest assessment of accuracy or other metrics. In practice, the test set includes input features and the correct answers, not the model’s outputs. The other statements are less precise descriptors: testing does reveal performance, but the crucial point about the data itself is that it holds the real answers and not the model’s predictions.