Which data concept is used to evaluate model performance on data the model has not seen?

Which data concept is used to evaluate model performance on data the model has not seen?

• A. Validation data
• B. Test data
• C. Training data
• D. Data drift

Answer: (B)

Focus: evaluating how a model performs on data it hasn’t seen before. The test data is set aside and kept separate from the training process, so it provides an unbiased measure of generalization—how the model would perform on new, real-world inputs. This final evaluation after training and tuning gives a truthful performance estimate because the model hasn’t been exposed to this data during learning. Validation data, while also unseen during training, is used during the development process to tune hyperparameters and monitor progress. Because it informs learning decisions, performance there can be optimistic if used to judge the final model. Data drift refers to changes in the data distribution over time, not to a data split used for evaluation. So, the data concept used to evaluate performance on data the model has not seen is the test data.

Focus: evaluating how a model performs on data it hasn’t seen before. The test data is set aside and kept separate from the training process, so it provides an unbiased measure of generalization—how the model would perform on new, real-world inputs. This final evaluation after training and tuning gives a truthful performance estimate because the model hasn’t been exposed to this data during learning.

Validation data, while also unseen during training, is used during the development process to tune hyperparameters and monitor progress. Because it informs learning decisions, performance there can be optimistic if used to judge the final model. Data drift refers to changes in the data distribution over time, not to a data split used for evaluation.

So, the data concept used to evaluate performance on data the model has not seen is the test data.