Skip to Main content Skip to Navigation

Utilisation des Divergences entre Mesures en Statistique Inférentielle

Abstract : We introduce estimation and test procedures through optimization of Divergence for discrete and continuous parametric models, for semiparametric two samples density ratio models and for nonparametric models restricted by linear constraints. The proposed procedures are based on a new dual representation for Divergences between measures. We show that the maximum parametric likelihood and the maximum empirical likelihood methods are particular cases corresponding to the choice of the modified Kullback-Leibler Divergence, and that the use of other Divergences leads to some estimates having similar properties even better in some cases. Several problems concerning the choice of the Divergence are noted for future investigations.
Document type :
Complete list of metadatas
Contributor : Amor Keziou <>
Submitted on : Tuesday, December 30, 2003 - 4:24:21 PM
Last modification on : Thursday, December 10, 2020 - 10:52:46 AM
Long-term archiving on: : Wednesday, September 12, 2012 - 12:15:36 PM


  • HAL Id : tel-00004069, version 1


Amor Keziou. Utilisation des Divergences entre Mesures en Statistique Inférentielle. Mathématiques [math]. Université Pierre et Marie Curie - Paris VI, 2003. Français. ⟨tel-00004069⟩



Record views


Files downloads