Tapered composite likelihood for spatial max-stable models

Handle URI:
http://hdl.handle.net/10754/566174
Title:
Tapered composite likelihood for spatial max-stable models
Authors:
Sang, Huiyan; Genton, Marc G. ( 0000-0001-6467-2998 )
Abstract:
Spatial extreme value analysis is useful to environmental studies, in which extreme value phenomena are of interest and meaningful spatial patterns can be discerned. Max-stable process models are able to describe such phenomena. This class of models is asymptotically justified to characterize the spatial dependence among extremes. However, likelihood inference is challenging for such models because their corresponding joint likelihood is unavailable and only bivariate or trivariate distributions are known. In this paper, we propose a tapered composite likelihood approach by utilizing lower dimensional marginal likelihoods for inference on parameters of various max-stable process models. We consider a weighting strategy based on a "taper range" to exclude distant pairs or triples. The "optimal taper range" is selected to maximize various measures of the Godambe information associated with the tapered composite likelihood function. This method substantially reduces the computational cost and improves the efficiency over equally weighted composite likelihood estimators. We illustrate its utility with simulation experiments and an analysis of rainfall data in Switzerland.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Publisher:
Elsevier BV
Journal:
Spatial Statistics
Issue Date:
May-2014
DOI:
10.1016/j.spasta.2013.07.003
Type:
Article
ISSN:
22116753
Sponsors:
The research of Huiyan Sang was partially supported by NSF grant DMS-1007618. Marc G. Genton's work was partially supported by DMS-1007504 and DMS-1100492. Both authors were supported by Award Number KUS-CI-016-04, from King Abdullah University of Science and Technology (KAUST). The authors thank Dr. Anthony Davison and Dr. Mathieu Ribatet for several useful discussions regarding this work and for providing the Swiss precipitation dataset.
Appears in Collections:
Articles; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorSang, Huiyanen
dc.contributor.authorGenton, Marc G.en
dc.date.accessioned2015-08-12T09:31:07Zen
dc.date.available2015-08-12T09:31:07Zen
dc.date.issued2014-05en
dc.identifier.issn22116753en
dc.identifier.doi10.1016/j.spasta.2013.07.003en
dc.identifier.urihttp://hdl.handle.net/10754/566174en
dc.description.abstractSpatial extreme value analysis is useful to environmental studies, in which extreme value phenomena are of interest and meaningful spatial patterns can be discerned. Max-stable process models are able to describe such phenomena. This class of models is asymptotically justified to characterize the spatial dependence among extremes. However, likelihood inference is challenging for such models because their corresponding joint likelihood is unavailable and only bivariate or trivariate distributions are known. In this paper, we propose a tapered composite likelihood approach by utilizing lower dimensional marginal likelihoods for inference on parameters of various max-stable process models. We consider a weighting strategy based on a "taper range" to exclude distant pairs or triples. The "optimal taper range" is selected to maximize various measures of the Godambe information associated with the tapered composite likelihood function. This method substantially reduces the computational cost and improves the efficiency over equally weighted composite likelihood estimators. We illustrate its utility with simulation experiments and an analysis of rainfall data in Switzerland.en
dc.description.sponsorshipThe research of Huiyan Sang was partially supported by NSF grant DMS-1007618. Marc G. Genton's work was partially supported by DMS-1007504 and DMS-1100492. Both authors were supported by Award Number KUS-CI-016-04, from King Abdullah University of Science and Technology (KAUST). The authors thank Dr. Anthony Davison and Dr. Mathieu Ribatet for several useful discussions regarding this work and for providing the Swiss precipitation dataset.en
dc.publisherElsevier BVen
dc.subjectComposite likelihooden
dc.subjectGeneralized extreme-value distributionen
dc.subjectMax-stable processen
dc.subjectStatistics of extremesen
dc.subjectWeighted composite likelihooden
dc.titleTapered composite likelihood for spatial max-stable modelsen
dc.typeArticleen
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.identifier.journalSpatial Statisticsen
dc.contributor.institutionDepartment of Statistics, Texas A and M University, College Station, TX 77843-3143, USAen
kaust.authorGenton, Marc G.en
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.