Hierarchical matrix approximation of large covariance matrices

Handle URI:
http://hdl.handle.net/10754/623623
Title:
Hierarchical matrix approximation of large covariance matrices
Authors:
Litvinenko, Alexander ( 0000-0001-5427-3598 ) ; Genton, Marc G. ( 0000-0001-6467-2998 ) ; Sun, Ying ( 0000-0001-6703-4270 )
Abstract:
We approximate large non-structured Matérn covariance matrices of size n×n in the H-matrix format with a log-linear computational cost and storage O(kn log n), where rank k ≪ n is a small integer. Applications are: spatial statistics, machine learning and image analysis, kriging and optimal design.
KAUST Department:
ECRC, KAUST; Spatio-temporal statistics & Data science, KAUST; Environmental Statistics Group, KAUST
Conference/Event name:
ECRC Advisory Board Meeting, KAUST
Issue Date:
30-Nov-2015
Type:
Poster
Sponsors:
ECRC, KAUST
Appears in Collections:
Posters

Full metadata record

DC FieldValue Language
dc.contributor.authorLitvinenko, Alexanderen
dc.contributor.authorGenton, Marc G.en
dc.contributor.authorSun, Yingen
dc.date.accessioned2017-05-16T08:21:21Z-
dc.date.available2017-05-16T08:21:21Z-
dc.date.issued2015-11-30-
dc.identifier.urihttp://hdl.handle.net/10754/623623-
dc.description.abstractWe approximate large non-structured Matérn covariance matrices of size n×n in the H-matrix format with a log-linear computational cost and storage O(kn log n), where rank k ≪ n is a small integer. Applications are: spatial statistics, machine learning and image analysis, kriging and optimal design.en
dc.description.sponsorshipECRC, KAUSTen
dc.subjectMatern covarianceen
dc.subjecthierarchical matricesen
dc.subjectdata compressionen
dc.subjectloglikelihood surrogateen
dc.subjectMLE methoden
dc.subjectParameter Estimationen
dc.titleHierarchical matrix approximation of large covariance matricesen
dc.typePosteren
dc.contributor.departmentECRC, KAUSTen
dc.contributor.departmentSpatio-temporal statistics & Data science, KAUSTen
dc.contributor.departmentEnvironmental Statistics Group, KAUSTen
dc.conference.dateNovember 2015en
dc.conference.nameECRC Advisory Board Meeting, KAUSTen
dc.conference.locationKAUST, B1en
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.