Glossary

Data leakage (ML) — Glossary

Improperly sharing information between training and evaluation that inflates metrics.

Reviewed:

Full definition

Leakage includes using future data, fitting scalers on the full dataset before splitting, or duplicates across train and test. It produces falsely optimistic deployment expectations.

Related encyclopedia topic

Cross-Validation — Cross-validation estimates how well a model generalizes by repeatedly training and evaluating on different splits of the data.

← Back to glossary · Learning hub