Cross-validation and a separate "supplied test set"

Question

After performing a 10 fold cross-validation, is it proper to perform a re-evaluation using a separate "supplied test set"?

Answer 1

Sure, why not. The more data sets you have, the better. The point of having separate data sets in the first place is to avoid overfitting of the models you learn to the data and getting a misleading impression of the performance of the model. Using separate test sets and techniques like cross-validation ensures that you get a more accurate and reasonable picture of the performance of your model.

The separate test set can be used to confirm the performance estimate computed during cross-validation.