Premium
The Shkarofsky‐Gneiting class of covariance models for bivariate Gaussian random fields
Author(s) -
Porcu Emilio,
Bevilacqua Moreno,
Hering Amanda S.
Publication year - 2018
Publication title -
stat
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.61
H-Index - 18
ISSN - 2049-1573
DOI - 10.1002/sta4.207
Subject(s) - bivariate analysis , covariance , covariance mapping , mathematics , covariance function , gaussian , parametric statistics , cauchy distribution , random field , parametric model , statistics , econometrics , covariance intersection , physics , quantum mechanics
We propose new covariance functions for bivariate Gaussian random fields that are very general and include as special cases other popular models proposed in earlier literature, namely, the bivariate Matérn and bivariate Cauchy models. The proposed model allows the covariance margins to belong to different parametric families with. To our knowledge, this is the first model of this type to be proposed in the literature. For instance, one of the margins can be of the Matérn type, whereas the latter can index long‐range dependence. Estimation of the model is illustrated through simulation.