Estimating the smoothing parameter in generalized spline-based regression: I. Cross-validatory criteria for binary data using small sample sizes (Q1861588): Difference between revisions

The author considers a nonparametric regression model in which the distribution of data is described by a one-parameter exponential family density and the estimated regression function links the covariates with the parameter of the family. It is supposed that this function is smooth enough. A spline-smoothing technique is used for estimation. Different methods of smoothing parameter validation are considered, such as full and robustified cross-validation. Results of simulations are presented for the estimation of different functions, e.g., \(h(s,t)=15(\exp(-s^2-t^2-st))-11\).