Neural Computation
October 1, 1996, Vol. 8, No. 7, Pages 1421-1426
(doi: 10.1162/neco.1996.8.7.1421)
No Free Lunch for Cross-Validation
Article PDF (285.25 KB)
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.