Test Data

Test data is a portion of the dataset held back and never used during training, used to measure how well a model performs on new, unseen examples.

Data & AI

Evaluating a model on the same data it trained on is misleading — it might have simply memorized the answers. The test set acts like a final exam with new questions, giving an honest estimate of real-world performance.

A common split is 80% training and 20% testing. Many workflows also use a third 'validation' set to tune settings while keeping the test set truly untouched until the very end.

Examples

Python

from sklearn.metrics import accuracy_score

predictions = model.predict(X_test)
print(accuracy_score(y_test, predictions))

Output:

0.92

Learn it in a lesson

See also

Back to the Dictionary