cross-validation

Evaluating models by training on subsets and testing on the rest

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

Etymology — From Latin `crux` (cross) + `validus` (strong, effective), from `valere` (to be strong)