Vector quantile regression beyond the specified case

Carlier, Guillaume; Chernozhukov, Victor; Galichon, Alfred

Carlier, Guillaume; Chernozhukov, Victor; Galichon, Alfred (2016), Vector quantile regression beyond the specified case, Journal of Multivariate Analysis, 161, p. 96-102. https://doi.org/10.1016/j.jmva.2017.07.003

Type

Article accepté pour publication ou publié

Lien vers un document non conservé dans cette base

https://arxiv.org/pdf/1610.06833.pdf

Date

2016-10

Nom de la revue

Journal of Multivariate Analysis

Volume

161

Éditeur

Academic Press

Pages

96-102

Identifiant publication

https://doi.org/10.1016/j.jmva.2017.07.003

Métadonnées

Afficher la notice complète

Auteur(s)

Carlier, Guillaume
CEntre de REcherches en MAthématiques de la DEcision [CEREMADE]
Chernozhukov, Victor
MIT Department of Economics
Galichon, Alfred
Courant Institute of Mathematical Sciences [New York] [CIMS]

Résumé (EN)

This paper studies vector quantile regression (VQR), which models the dependence of a random vector with respect to a vector of explanatory variables with enough flexibility to capture the whole conditional distribution, and not only the conditional mean. The problem of vector quantile regression is formulated as an optimal transport problem subject to an additional mean-independence condition. This paper provides results on VQR beyond the specified case which had been the focus of previous work. We show that even beyond the specified case, the VQR problem still has a solution which provides a general representation of the conditional dependence between random vectors.

Mots-clés

Duality; Optimal transport; Vector quantile regression