Cross-Validation

/ˌkrɔːs ˌvælɪˈdeɪʃən/cross·val·i·da·tionnoun
Model Evaluation CoreFoundational

Definition

1.[in machine learning] a technique used to estimate the performance of a model on unseen data by partitioning the dataset into multiple subsets.

We used k-fold Cross-Validation to ensure our model's performance was robust across different data splits.

Formal statement

Etymology

From cross- 'crosswise, reciprocally' plus validation, Latin validus 'strong'. The prefix names the reciprocal arrangement in which each subset serves in turn as the held-out set. Resampling of this kind appears in the 1930s; the method was formalised and named by Stone and by Geisser, independently, in 1974.

Synonyms

  • k-fold validationsense 1 · Near

See also

References

  • Hastie, T., Tibshirani, R., & Friedman, J. (2009). The Elements of Statistical Learning, 2nd ed., ch. 5.Springer.