La plateforme OSKAR Bordeaux évolue pour rejoindre l'archive ouverte HAL. Retrouvez tous vos dépôts sur le nouveau portail HAL UB : https://u-bordeaux.hal.science/. Pour toute aide ou information, contactez-nous info@oskar-bordeaux.fr
STOCHASTIC APPROXIMATION ALGORITHMS FOR SUPERQUANTILES ESTIMATION
Langue
en
Document de travail - Pré-publication
Résumé en anglais
This paper is devoted to two different two-timescale stochastic approximation algorithms for superquantile estimation. We shall investigate the asymptotic behavior of a Robbins-Monro estimator and its convexified version. ...Lire la suite >
This paper is devoted to two different two-timescale stochastic approximation algorithms for superquantile estimation. We shall investigate the asymptotic behavior of a Robbins-Monro estimator and its convexified version. Our main contribution is to establish the almost sure convergence, the quadratic strong law and the law of iterated logarithm for our estimates via a martingale approach. A joint asymptotic normality is also provided. Our theoretical analysis is illustrated by numerical experiments on real datasets.< Réduire
Origine
Importé de halUnités de recherche