A Bernstein-polynomial sieve maximum likelihood estimator for isotropic covariance functions valid in R^∞ is shown to be consistent under increasing-domain asymptotics and to achieve lower L∞ and L2 errors than parametric or other nonparametric alternatives in simulations.
(2016), On the smallest eigenvalues of covariance matrices of multivariate spatial processes , Stat, 1, 102--107
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Nonparametric Estimation of Isotropic Covariance Function
A Bernstein-polynomial sieve maximum likelihood estimator for isotropic covariance functions valid in R^∞ is shown to be consistent under increasing-domain asymptotics and to achieve lower L∞ and L2 errors than parametric or other nonparametric alternatives in simulations.