Robust cross-validation score function for non-linear function estimation

Abstract

In this paper a new method for tuning regularisation parameters or other hyperparameters of a learning process (non-linear function estimation) is proposed, called robust cross-validation score function (CV S-foldRobust). CV S-foldRobust is effective for dealing with outliers and non-Gaussian noise distributions on the data. Illustrative simulation results are given to demonstrate that the CV S-foldRobustmethod outperforms other cross-validation methods. © Springer-Verlag Berlin Heidelberg 2002.