Concentration inequalities of the cross-validation estimator for empirical risk minimizer

We derive concentration inequalities for the cross-validation estimate of the generalization error for empirical risk minimizers. In the general setting, we show that the worst-case error of this estimate is not much worse that of training error estimate see Kearns M, Ron D. [Algorithmic stability and sanity-check bounds for leave-one-out cross-validation. Neural Comput. 1999;11:1427-1453]. General loss functions and class of predictors with finite VC-dimension are considered. Our focus is on proving the consistency of the various cross-validation procedures. We point out the interest of each cross-validation procedure in terms of rates of convergence. An interesting consequence is that the size of the test sample is not required to grow to infinity for the consistency of the cross-validation procedure.