Show simple item record

dc.contributor.authorDalalyan, Arnak S.*
dc.contributor.authorHebiri, Mohamed*
dc.contributor.authorMeziani, Katia*
dc.contributor.authorSalmon, Joseph*
dc.date.accessioned2015-04-14T17:39:33Z
dc.date.available2015-04-14T17:39:33Z
dc.date.issued2013
dc.identifier.urihttps://basepub.dauphine.fr/handle/123456789/14943
dc.language.isoenen
dc.subjectMachine Learningen
dc.subject.ddc519en
dc.titleLearning Heteroscedastic Models by Convex Programming under Group Sparsityen
dc.typeCommunication / Conférence
dc.description.abstractenPopular sparse estimation methods based on ℓ1-relaxation, such as the Lasso and the Dantzig selector, require the knowledge of the variance of the noise in order to properly tune the regularization parameter. This constitutes a major obstacle in applying these methods in several frameworks---such as time series, random fields, inverse problems---for which the noise is rarely homoscedastic and its level is hard to know in advance. In this paper, we propose a new approach to the joint estimation of the conditional mean and the conditional variance in a high-dimensional (auto-) regression setting. An attractive feature of the proposed estimator is that it is efficiently computable even for very large scale problems by solving a second-order cone program (SOCP). We present theoretical analysis and numerical results assessing the performance of the proposed procedure.en
dc.identifier.citationpages379-387en
dc.relation.ispartofseriesnumber28en
dc.relation.ispartoftitleProceedings of The 30th International Conference on Machine Learningen
dc.relation.ispartofeditorDasgupta, Sanjoy
dc.relation.ispartofeditorMcAllester, David
dc.relation.ispartofdate2013
dc.subject.ddclabelProbabilités et mathématiques appliquéesen
dc.relation.conftitle30th International Conference on Machine Learningen
dc.relation.confdate2013-06
dc.relation.confcityAtlantaen
dc.relation.confcountryÉtats-Unisen
dc.relation.forthcomingnonen
hal.person.labIds*
hal.person.labIds*
hal.person.labIds*
hal.person.labIds*


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record