Glossary · Phase 3 · Machine learning
Cross-validation
Scoring a model several times on different held-back slices of the data, for a steadier estimate.
Try it now
Every row gets to be test data once, so one lucky split can't fool you.
Learn it properly
Cross-validation · Reuse limited data for more reliable estimates. About 30 minutes.
Other terms from the same lesson
- fold: One of the equal slices that cross-validation cuts the data into. Each fold takes a turn as the held-back data.
- grid search: Trying every combination from a list of settings, scoring each with cross-validation, and keeping the best.
- hyperparameter: A setting you choose before training, such as tree depth or learning rate, rather than one the model learns.
Never trained an AI?
Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account.