A second-order method for convex L1-regularized optimization with active-set prediction

Basit Öğe Kaydı
dc.authorwosidNocedal, Jorge/B-7255-2009
dc.authorwosidWaechter, Andreas/G-2499-2011
dc.authorwosidOztoprak, Figen/ABH-1969-2021
dc.contributor.authorKeskar, N.
dc.contributor.authorNocedal, J.
dc.contributor.authorOztoprak, F.
dc.contributor.authorWaechter, A.
dc.date.accessioned2024-07-18T20:45:12Z
dc.date.available2024-07-18T20:45:12Z
dc.date.issued2016
dc.departmentİstanbul Bilgi Üniversitesien_US
dc.description.abstractWe describe an active-set method for the minimization of an objective function phi that is the sum of a smooth convex function f and an l(1)-regularization term. A distinctive feature of the method is the way in which active-set identification and second-order subspace minimization steps are integrated to combine the predictive power of the two approaches. At every iteration, the algorithm selects a candidate set of free and fixed variables, performs an (inexact) subspace phase, and then assesses the quality of the new active set. If it is not judged to be acceptable, then the set of free variables is restricted and a new active-set prediction is made. We establish global convergence for our approach under the assumptions of Lipschitz-continuity and strong-convexity of f, and compare the new method against state-of-the-art codes.en_US
dc.description.sponsorshipDirect For Mathematical & Physical Scien; Division Of Mathematical Sciences [1216567] Funding Source: National Science Foundationen_US
dc.identifier.doi10.1080/10556788.2016.1138222
dc.identifier.endpage621en_US
dc.identifier.issn1055-6788
dc.identifier.issn1029-4937
dc.identifier.issue3en_US
dc.identifier.scopus2-s2.0-84991401462en_US
dc.identifier.scopusqualityQ1en_US
dc.identifier.startpage605en_US
dc.identifier.urihttps://doi.org/10.1080/10556788.2016.1138222
dc.identifier.urihttps://hdl.handle.net/11411/7447
dc.identifier.volume31en_US
dc.identifier.wosWOS:000374781100011en_US
dc.identifier.wosqualityQ2en_US
dc.indekslendigikaynakWeb of Scienceen_US
dc.indekslendigikaynakScopusen_US
dc.language.isoenen_US
dc.publisherTaylor & Francis Ltden_US
dc.relation.ispartofOptimization Methods & Softwareen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/closedAccessen_US
dc.subjectL(1)-Minimizationen_US
dc.subjectSecond-Orderen_US
dc.subjectActive-Set Predictionen_US
dc.subjectActive-Set Correctionen_US
dc.subjectSubspace-Optimizationen_US
dc.subject49men_US
dc.subject65ken_US
dc.subject65hen_US
dc.subject90cen_US
dc.subjectLinear Inverse Problemsen_US
dc.subjectThresholding Algorithmen_US
dc.subjectLogistic-Regressionen_US
dc.subjectCoordinate Descenten_US
dc.subjectMinimizationen_US
dc.subjectShrinkageen_US
dc.titleA second-order method for convex L1-regularized optimization with active-set predictionen_US
dc.typeArticleen_US

Dosyalar

Koleksiyon

Web of Science Indexed Publications
Scopus Indexed Publications