Vecchia approximations of Gaussian-process predictions
Gaussian processes (GPs) are highly flexible function estimators used for geospatial analysis, nonparametric regression, and machine learning, but they are computationally infeasible for large datasets. Vecchia approximations of GPs have been used to enable fast evaluation of the likelihood for para...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2020-05 |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Katzfuss, Matthias Guinness, Joseph Gong, Wenlong Zilber, Daniel |
description | Gaussian processes (GPs) are highly flexible function estimators used for geospatial analysis, nonparametric regression, and machine learning, but they are computationally infeasible for large datasets. Vecchia approximations of GPs have been used to enable fast evaluation of the likelihood for parameter inference. Here, we study Vecchia approximations of spatial predictions at observed and unobserved locations, including obtaining joint predictive distributions at large sets of locations. We consider a general Vecchia framework for GP predictions, which contains some novel and some existing special cases. We study the accuracy and computational properties of these approaches theoretically and numerically, proving that our new methods exhibit linear computational complexity in the total number of spatial locations. We show that certain choices within the framework can have a strong effect on uncertainty quantification and computational cost, which leads to specific recommendations on which methods are most suitable for various settings. We also apply our methods to a satellite dataset of chlorophyll fluorescence, showing that the new methods are faster or more accurate than existing methods, and reduce unrealistic artifacts in prediction maps. |
doi_str_mv | 10.48550/arxiv.1805.03309 |
format | Article |
fullrecord | <record><control><sourceid>proquest_arxiv</sourceid><recordid>TN_cdi_arxiv_primary_1805_03309</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2073317491</sourcerecordid><originalsourceid>FETCH-LOGICAL-a521-5dcb9ddc4c401ab14646f4e2a83a22d61a15f296f108916232ae147167519db23</originalsourceid><addsrcrecordid>eNotj01Lw0AURQdBsNT-AFcGXCfOe_ORzFKKVqHgprgNLzMTnKJJnGmk_nvH1tVd3MvlHMZugFeyUYrfUzyG7woariouBDcXbIFCQNlIxCu2SmnPOUddo1JiwfSbt_Y9UEHTFMdj-KRDGIdUjH2xoTmlQEOZC-tTKqboXbCn_ppd9vSR_Oo_l2z39LhbP5fb183L-mFbkkIolbOdcc5KKzlQB1JL3UuP1AhCdBoIVI9G98AbAxoFkgdZg64VGNehWLLb8-1Jqp1i5os_7Z9ce5LLi7vzIkN-zT4d2v04xyEztcjr7F1LA-IXbNhQ-w</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2073317491</pqid></control><display><type>article</type><title>Vecchia approximations of Gaussian-process predictions</title><source>arXiv.org</source><source>Free E- Journals</source><creator>Katzfuss, Matthias ; Guinness, Joseph ; Gong, Wenlong ; Zilber, Daniel</creator><creatorcontrib>Katzfuss, Matthias ; Guinness, Joseph ; Gong, Wenlong ; Zilber, Daniel</creatorcontrib><description>Gaussian processes (GPs) are highly flexible function estimators used for geospatial analysis, nonparametric regression, and machine learning, but they are computationally infeasible for large datasets. Vecchia approximations of GPs have been used to enable fast evaluation of the likelihood for parameter inference. Here, we study Vecchia approximations of spatial predictions at observed and unobserved locations, including obtaining joint predictive distributions at large sets of locations. We consider a general Vecchia framework for GP predictions, which contains some novel and some existing special cases. We study the accuracy and computational properties of these approaches theoretically and numerically, proving that our new methods exhibit linear computational complexity in the total number of spatial locations. We show that certain choices within the framework can have a strong effect on uncertainty quantification and computational cost, which leads to specific recommendations on which methods are most suitable for various settings. We also apply our methods to a satellite dataset of chlorophyll fluorescence, showing that the new methods are faster or more accurate than existing methods, and reduce unrealistic artifacts in prediction maps.</description><identifier>EISSN: 2331-8422</identifier><identifier>DOI: 10.48550/arxiv.1805.03309</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Approximation ; Chlorophyll ; Computation ; Fluorescence ; Gaussian process ; Geographic information systems ; Machine learning ; Predictions ; Regression analysis ; Spatial analysis ; Statistics - Computation ; Statistics - Methodology</subject><ispartof>arXiv.org, 2020-05</ispartof><rights>2020. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,784,885,27925</link.rule.ids><backlink>$$Uhttps://doi.org/10.1007/s13253-020-00401-7$$DView published paper (Access to full text may be restricted)$$Hfree_for_read</backlink><backlink>$$Uhttps://doi.org/10.48550/arXiv.1805.03309$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Katzfuss, Matthias</creatorcontrib><creatorcontrib>Guinness, Joseph</creatorcontrib><creatorcontrib>Gong, Wenlong</creatorcontrib><creatorcontrib>Zilber, Daniel</creatorcontrib><title>Vecchia approximations of Gaussian-process predictions</title><title>arXiv.org</title><description>Gaussian processes (GPs) are highly flexible function estimators used for geospatial analysis, nonparametric regression, and machine learning, but they are computationally infeasible for large datasets. Vecchia approximations of GPs have been used to enable fast evaluation of the likelihood for parameter inference. Here, we study Vecchia approximations of spatial predictions at observed and unobserved locations, including obtaining joint predictive distributions at large sets of locations. We consider a general Vecchia framework for GP predictions, which contains some novel and some existing special cases. We study the accuracy and computational properties of these approaches theoretically and numerically, proving that our new methods exhibit linear computational complexity in the total number of spatial locations. We show that certain choices within the framework can have a strong effect on uncertainty quantification and computational cost, which leads to specific recommendations on which methods are most suitable for various settings. We also apply our methods to a satellite dataset of chlorophyll fluorescence, showing that the new methods are faster or more accurate than existing methods, and reduce unrealistic artifacts in prediction maps.</description><subject>Approximation</subject><subject>Chlorophyll</subject><subject>Computation</subject><subject>Fluorescence</subject><subject>Gaussian process</subject><subject>Geographic information systems</subject><subject>Machine learning</subject><subject>Predictions</subject><subject>Regression analysis</subject><subject>Spatial analysis</subject><subject>Statistics - Computation</subject><subject>Statistics - Methodology</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><sourceid>GOX</sourceid><recordid>eNotj01Lw0AURQdBsNT-AFcGXCfOe_ORzFKKVqHgprgNLzMTnKJJnGmk_nvH1tVd3MvlHMZugFeyUYrfUzyG7woariouBDcXbIFCQNlIxCu2SmnPOUddo1JiwfSbt_Y9UEHTFMdj-KRDGIdUjH2xoTmlQEOZC-tTKqboXbCn_ppd9vSR_Oo_l2z39LhbP5fb183L-mFbkkIolbOdcc5KKzlQB1JL3UuP1AhCdBoIVI9G98AbAxoFkgdZg64VGNehWLLb8-1Jqp1i5os_7Z9ce5LLi7vzIkN-zT4d2v04xyEztcjr7F1LA-IXbNhQ-w</recordid><startdate>20200514</startdate><enddate>20200514</enddate><creator>Katzfuss, Matthias</creator><creator>Guinness, Joseph</creator><creator>Gong, Wenlong</creator><creator>Zilber, Daniel</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope><scope>EPD</scope><scope>GOX</scope></search><sort><creationdate>20200514</creationdate><title>Vecchia approximations of Gaussian-process predictions</title><author>Katzfuss, Matthias ; Guinness, Joseph ; Gong, Wenlong ; Zilber, Daniel</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a521-5dcb9ddc4c401ab14646f4e2a83a22d61a15f296f108916232ae147167519db23</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Approximation</topic><topic>Chlorophyll</topic><topic>Computation</topic><topic>Fluorescence</topic><topic>Gaussian process</topic><topic>Geographic information systems</topic><topic>Machine learning</topic><topic>Predictions</topic><topic>Regression analysis</topic><topic>Spatial analysis</topic><topic>Statistics - Computation</topic><topic>Statistics - Methodology</topic><toplevel>online_resources</toplevel><creatorcontrib>Katzfuss, Matthias</creatorcontrib><creatorcontrib>Guinness, Joseph</creatorcontrib><creatorcontrib>Gong, Wenlong</creatorcontrib><creatorcontrib>Zilber, Daniel</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection><collection>arXiv Statistics</collection><collection>arXiv.org</collection><jtitle>arXiv.org</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Katzfuss, Matthias</au><au>Guinness, Joseph</au><au>Gong, Wenlong</au><au>Zilber, Daniel</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Vecchia approximations of Gaussian-process predictions</atitle><jtitle>arXiv.org</jtitle><date>2020-05-14</date><risdate>2020</risdate><eissn>2331-8422</eissn><abstract>Gaussian processes (GPs) are highly flexible function estimators used for geospatial analysis, nonparametric regression, and machine learning, but they are computationally infeasible for large datasets. Vecchia approximations of GPs have been used to enable fast evaluation of the likelihood for parameter inference. Here, we study Vecchia approximations of spatial predictions at observed and unobserved locations, including obtaining joint predictive distributions at large sets of locations. We consider a general Vecchia framework for GP predictions, which contains some novel and some existing special cases. We study the accuracy and computational properties of these approaches theoretically and numerically, proving that our new methods exhibit linear computational complexity in the total number of spatial locations. We show that certain choices within the framework can have a strong effect on uncertainty quantification and computational cost, which leads to specific recommendations on which methods are most suitable for various settings. We also apply our methods to a satellite dataset of chlorophyll fluorescence, showing that the new methods are faster or more accurate than existing methods, and reduce unrealistic artifacts in prediction maps.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><doi>10.48550/arxiv.1805.03309</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2020-05 |
issn | 2331-8422 |
language | eng |
recordid | cdi_arxiv_primary_1805_03309 |
source | arXiv.org; Free E- Journals |
subjects | Approximation Chlorophyll Computation Fluorescence Gaussian process Geographic information systems Machine learning Predictions Regression analysis Spatial analysis Statistics - Computation Statistics - Methodology |
title | Vecchia approximations of Gaussian-process predictions |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-04T23%3A05%3A17IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_arxiv&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Vecchia%20approximations%20of%20Gaussian-process%20predictions&rft.jtitle=arXiv.org&rft.au=Katzfuss,%20Matthias&rft.date=2020-05-14&rft.eissn=2331-8422&rft_id=info:doi/10.48550/arxiv.1805.03309&rft_dat=%3Cproquest_arxiv%3E2073317491%3C/proquest_arxiv%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2073317491&rft_id=info:pmid/&rfr_iscdi=true |