Hierarchical matrix approximation of large covariance matrices

dc.contributor.author	Litvinenko, Alexander
dc.contributor.author	Genton, Marc G.
dc.contributor.author	Sun, Ying
dc.date.accessioned	2017-05-16T08:21:21Z
dc.date.available	2017-05-16T08:21:21Z
dc.date.issued	2015-11-30
dc.identifier.uri	http://hdl.handle.net/10754/623623
dc.description.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.
dc.description.sponsorship	ECRC, KAUST
dc.subject	Matern covariance
dc.subject	hierarchical matrices
dc.subject	data compression
dc.subject	loglikelihood surrogate
dc.subject	MLE method
dc.subject	Parameter Estimation
dc.title	Hierarchical matrix approximation of large covariance matrices
dc.type	Poster
dc.contributor.department	ECRC, KAUST
dc.contributor.department	Spatio-temporal statistics & Data science, KAUST
dc.contributor.department	Environmental Statistics Group, KAUST
dc.conference.date	November 2015
dc.conference.name	ECRC Advisory Board Meeting, KAUST
dc.conference.location	KAUST, B1
refterms.dateFOA	2018-06-13T15:30:35Z

Name:

poster_litvinenko_adv_board.pdf

Size:

548.5Kb

Format:

PDF

Description:

Identification of hyper-parameters of Matern covariance function via Hierarchical matrix technique