Skip to Main content Skip to Navigation

Validation croisée et pénalisation pour l'estimation de densité

Abstract : This thesis takes place in the density estimation setting from a nonparametric and nonasymptotic point of view. It concerns the statistical algorithm selection problem which generalizes, among others, the problem of model and bandwidth selection. We study classical procedures, such as penalization or resampling procedures (in particular V-fold cross-validation), which evaluate an algorithm by estimating its risk. We provide, thanks to concentration inequalities, an optimal penalty for selecting a linear estimator and we prove oracle inequalities and adaptative properties for resampling procedures. Moreover, new resampling procedure, based on estimator comparison by the mean of robust tests, is introduced as an alternative to procedures relying on the unbiased risk estimation principle. A second goal of this work is to compare these procedures from a theoretical point of view and to understand the role of V for V-fold penalization. We validate these theoretical results on empirical studies.
Complete list of metadata

Cited literature [169 references]  Display  Hide  Download
Contributor : ABES STAR :  Contact
Submitted on : Wednesday, June 17, 2015 - 11:57:05 AM
Last modification on : Saturday, June 25, 2022 - 9:09:34 PM
Long-term archiving on: : Tuesday, September 15, 2015 - 5:46:08 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01164581, version 1


Nelo Molter Magalhães Magalhães. Validation croisée et pénalisation pour l'estimation de densité. Probabilités [math.PR]. Université Paris Sud - Paris XI, 2015. Français. ⟨NNT : 2015PA112100⟩. ⟨tel-01164581⟩



Record views


Files downloads