Skip to content

Spatial regression model

2 messages · Nadine Henkenjohann, Roger Bivand

#
Hello,

is there a package which allows to fit a spatial regression model  for 
any k-dimensional data?
The spatial regression model is composed of three components:

y=mu+Z+epsilon,   

where mu is a low order polynomial, Z is a realization of a second-order 
stochastic process with covariance matrix Sigma and epsilon is a 
realization of a white-noise process. The covariance function belongs to 
the class of  either isotropic or anisotropic functions. Mean and 
covariance parameters are estimated simultaneously by ML or REML.

I hope somebody can answer my question.
Thank you!

Nadine Henkenjohann
#
On Wed, 7 Jan 2004, Nadine Henkenjohann wrote:

            
This site:

http://sal.agecon.uiuc.edu/csiss/Rgeo/index.html

tries to provide some guidance about the availability of packages for 
analysis of spatial data. The Geostatistics section of that site lists a 
number of packages perhaps relevant to your problem, but there are other 
possibilities too.

Roger