cross-validation

Go back to the [[AI Glossary]]

A mechanism for estimating how well a model will generalize to new data by testing the model against one or more non-overlapping data subsets withheld from the training set.