Isotropic Covariance Matrix Polynomials on Spheres

Chunsheng Ma
DOI: https://doi.org/10.1080/07362994.2016.1170612
2016-01-01
Stochastic Analysis and Applications
Abstract:This article characterizes the covariance matrix function of a Gaussian or elliptically contoured vector random field that is stationary, isotropic, and mean square continuous on the sphere. By applying the characterization to examine the validity of a matrix function whose entries are polynomials of degrees up to 4, we obtain a necessary and sufficient condition for the polynomial matrix to be an isotropic covariance matrix function on the sphere.
What problem does this paper attempt to address?