[R-sig-Geo] Cross validation of ST Kriging
dawn.williams.10 at ucl.ac.uk
Fri Aug 21 16:46:08 CEST 2015
I would like to conduct cross validation on ST kriging performed with different ST variograms. I understand, that it is not yet possible to conduct cross validation on ST Kriging but that one should assume an extremely local temporal neighbourhood and perform cross validation on spatial interpolation. My question is how do I reflect the various space-time covariance structures in doing this if the krige.cv function will only accept a vgm? How do I input the other parameters of the ST covariance models in creating the vgm for comparison?
Thank you in advance.
[[alternative HTML version deleted]]
More information about the R-sig-Geo