A class of additive covariance models of an isotropic random process is proposed, motivated by the spectral representation of the covariance function. Model parameters are estimated by using a special case of the minimum norm quadratic estimation estimator, whose asymptotic moments have convenient expressions in terms of spectral densities. Fitting a model in this class is equivalent to fitting an additive model of the spectral density. The class of spectral additive models proposed is dense in the set of summable covariance functions having a spectral density, allowing approximately unbiased estimation of an arbitrary covariance function and its spectral density. Theoretical results are supported by numerical comparison with commonly used models. A procedure to assist model selection is proposed. The techniques are illustrated with an application to contaminant data. Copyright (c) 2008 Royal Statistical Society.
Download Info
To download:
If you experience problems downloading a file, check if you have the
proper application to
view it first. Information about this may be contained
in the File-Format links below. In case of further problems read
the IDEAS help
page. Note that these files are not on the IDEAS
site. Please be patient as the files may be large.
As the access to this document is restricted, you may want to look for a different version under "Related research" (further below) or search for a different version of it.