Maxisets for linear procedures

Vincent Rivoirard (1)

1 LPMA - Laboratoire de Probabilités et Modèles Aléatoires

We study maxisets for linear procedures in the framework of the heteroscedastic white noise model. This enables us to point out the nature of the spaces naturally connected to these procedures and to compare the performance of linear and non linear estimates by comparing their respective maxisets.

Mots clés: Besov spaces; Heteroscedastic white noise model; Linear estimators; Maxisets; Rate of convergence

Domaines: Statistiques [math.ST]; Théorie [stat.TH]

Vincent Rivoirard. Maxisets for linear procedures. Statistics and Probability Letters, 2004, 67 (3), pp.267-275. ⟨10.1016/j.spl.2003.12.010⟩. ⟨hal-00634272⟩