Empirical L2-distance between regression model and vector space
Résumé
The paper is devoted to the estimation of the square distance between the régression function / and the subspace Ep spanned by the linearly independent functions gi,...,gp. In order to estimate this measure of discreapancy, we propose an empirical L2-distance between / and Ep, without weight and we show that it is invariant with respect to change of basis in Ep.
Domaines
Statistiques [math.ST]| Origine | Accord explicite pour ce dépôt |
|---|