This paper1 advances results in model selection by relaxing the task of optimally tuning the regularization parameter in a number of algorithms with respect to the classical cross-validation performance criterion as a convex optimization problem. The proposed strategy differs from the scope of e.g. generalized cross-validation (GCV) as it concerns the efficient optimization, not the individual ...

Since its introduction by Stone (1974) and Geisser (1975), cross-validation has been studied and improved by several authors including Burman et al. the most widely used and best known is generalized cross-validation (GCV) (Craven & Wahba, 1979), which establishes a single-pass method that penalizes the fit by the trace of the smoother matrix assuming independent errors. We propose an extension...

Smoothing splines with generalized cross-validation parameter selection (GCV-spline) provide a method to find an optimal smoother for an fMRI time series. The purpose of this study was to compare the variance of parameter estimates and the bias of the variance estimator for a linear regression model smoothed with GCV-spline and the low-pass filter in SPM99 (SPM-HRF). The mean bias with the SPM-...

The point spread function (PSF) of a blurred image is often unknown a priori; the blur must first be identified from the degraded image data before restoring the image. Generalized cross-validation (GCV) is introduced to address the blur identification problem. The GCV criterion identifies model parameters for the blur, the image, and the regularization parameter, providing all the information ...

This paper considers optimization of the ridge parameters in generalized ridge regression (GRR) by minimizing a model selection criterion. GRR has a major advantage over ridge regression (RR) in that a solution to the minimization problem for one model selection criterion, i.e., Mallows’ Cp criterion, can be obtained explicitly with GRR, but such a solution for any model selection criteria, e.g...

This article considers spline smoothing of variance functions. We focus on selection of smoothing parameters and develop three direct data-driven methods: unbiased risk (UBR), generalized approximate cross validation (GACV) and generalized maximum likelihood (GML). In addition to guaranteed convergence, simulations show that these direct methods perform better than existing indirect UBR, genera...

An extension of reproducing kernel Hilbert space (RKHS) theory provides a new framework for modeling functional regression models with functional responses. The approach only presumes a general nonlinear regression structure as opposed to previously studied linear regression models. Generalized cross-validation (GCV) is proposed for automatic smoothing parameter estimation. The new RKHS estimat...

In this paper we study advantages and limitations of the Generalized Cross Validation (GCV) approach for selecting a regularization parameter in the case of a partially stochastic linear irregular operator equation. The research has been motivated by an inverse problem in epidemiology, where the goal was to reconstruct a time dependent treatment recovery rate for Plasmodium falciparum, the most...

| Superresolution reconstruction produces a high resolution image from a set of aliased low resolution images. We model the low resolution frames as blurred and down-sampled, shifted versions of the high resolution image. In many applications, the blurring process, i.e., point spread function (PSF) parameters of the imaging system, is not known. In our blind superresolution algorithm, we rst es...