A comparison between Markov approximations and other methods for large spatial data sets
(2013) In Computational Statistics & Data Analysis 61. p.7-21- Abstract
- The Matern covariance function is a popular choice for modeling dependence in spatial environmental data. Standard Matern covariance models are, however, often computationally infeasible for large data sets. Recent results for Markov approximations of Gaussian Matern fields based on Hilbert space approximations are extended using wavelet basis functions. Using a simulation-based study, these Markov approximations are compared with two of the most popular methods for computationally efficient model approximations, covariance tapering and the process convolution method. The methods are compared with respect to their computational properties when used for spatial prediction (kriging), and the results show that, for a given computational cost,... (More)
- The Matern covariance function is a popular choice for modeling dependence in spatial environmental data. Standard Matern covariance models are, however, often computationally infeasible for large data sets. Recent results for Markov approximations of Gaussian Matern fields based on Hilbert space approximations are extended using wavelet basis functions. Using a simulation-based study, these Markov approximations are compared with two of the most popular methods for computationally efficient model approximations, covariance tapering and the process convolution method. The methods are compared with respect to their computational properties when used for spatial prediction (kriging), and the results show that, for a given computational cost, the Markov methods have a substantial gain in accuracy compared with the other methods. (C) 2012 Elsevier B.V. All rights reserved. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/3669838
- author
- Bolin, David LU and Lindgren, Finn LU
- organization
- publishing date
- 2013
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Matern covariances, Kriging, Wavelets, Markov random fields, Covariance, tapering, Process convolutions
- in
- Computational Statistics & Data Analysis
- volume
- 61
- pages
- 7 - 21
- publisher
- Elsevier
- external identifiers
-
- wos:000315552600002
- scopus:84885018951
- ISSN
- 0167-9473
- DOI
- 10.1016/j.csda.2012.11.011
- language
- English
- LU publication?
- yes
- id
- 46af3df2-01dd-4de1-9391-a001bc316d5f (old id 3669838)
- date added to LUP
- 2016-04-01 10:48:46
- date last changed
- 2022-04-28 01:39:58
@article{46af3df2-01dd-4de1-9391-a001bc316d5f, abstract = {{The Matern covariance function is a popular choice for modeling dependence in spatial environmental data. Standard Matern covariance models are, however, often computationally infeasible for large data sets. Recent results for Markov approximations of Gaussian Matern fields based on Hilbert space approximations are extended using wavelet basis functions. Using a simulation-based study, these Markov approximations are compared with two of the most popular methods for computationally efficient model approximations, covariance tapering and the process convolution method. The methods are compared with respect to their computational properties when used for spatial prediction (kriging), and the results show that, for a given computational cost, the Markov methods have a substantial gain in accuracy compared with the other methods. (C) 2012 Elsevier B.V. All rights reserved.}}, author = {{Bolin, David and Lindgren, Finn}}, issn = {{0167-9473}}, keywords = {{Matern covariances; Kriging; Wavelets; Markov random fields; Covariance; tapering; Process convolutions}}, language = {{eng}}, pages = {{7--21}}, publisher = {{Elsevier}}, series = {{Computational Statistics & Data Analysis}}, title = {{A comparison between Markov approximations and other methods for large spatial data sets}}, url = {{http://dx.doi.org/10.1016/j.csda.2012.11.011}}, doi = {{10.1016/j.csda.2012.11.011}}, volume = {{61}}, year = {{2013}}, }