cross-validation

Evaluating models by training on subsets and testing on the rest

“Cross-validation revealed the model's true generalization performance.”

Origin: From Latin crux (cross) + validus (strong, effective), from valere (to be strong)