No Free Lunch for Cross-Validation
- 1 October 1996
- journal article
- Published by MIT Press in Neural Computation
- Vol. 8 (7) , 1421-1426
- https://doi.org/10.1162/neco.1996.8.7.1421
Abstract
It is known theoretically that an algorithm cannot be good for an arbitrary prior. We show that in practical terms this also applies to the technique of “cross-validation,” which has been widely regarded as defying this general rule. Numerical examples are analyzed in detail. Their implications to researches on learning algorithms are discussed.Keywords
This publication has 1 reference indexed in Scilit:
- Theory of Statistical EstimationMathematical Proceedings of the Cambridge Philosophical Society, 1925