[R] kriging in R

[Ernesto Jardim]

Hi

If you read the description of "varcov.spatial" you'll see that it is
used to *predict* a covariance matrix, based on the parameters of the
covariance function. So you don't need the oebserved data, you need
parameters for the covariance function.

Regards

EJ

On Thu, 2003-04-17 at 05:43, Yan Yu wrote:
> THanks for the suggestion.
> I have a Q when trying to use it..
> I want to get covariance matrix from the data set in hand..
> basically i have a rainfall data set in a 15x12 points.
> i try to use:
> "varcov.spatial", but it seems that the input to that function is just
> coordinate list of the data values, not the actual data at those points.
> I am confused by this?
>  How come the covariance matrix or variogram matrix is irrelevant to the
> the data value, but only depends on the data configuration(i.e., the
> sampling locations)?
> I found other covariance functions, like cor.number or cov.spatial is
> similar in the sense that they only ask for data locations, not data values..
> 
> Any hints or pointer is appreciated!
> yan
> 
> 
> On 17 Jan 2003, Ernesto Jardim wrote:
> 
> > Hi
> >
> > I'm working with geostatistics and I use geoR package, which I find very
> > good. However it doesn't implement the classic geostatistics but a
> > "model based geostatistics" as proposed by Peter Diggle and team.
> >
> > R news of June 2001 has an article about geoR by Paulo Ribeiro, the main
> > developer, that you might like to take a look at. I find this number of
> > Rnews a very good source for someone willing to use geostatistics in R.
> >
> > Regards
> >
> > EJ
> >
> > On Fri, 2003-01-17 at 05:06, Yan Yu wrote:
> > > Hi, all,
> > >    Have anyone used kringing included in R?  How is it?
> > > Does it handle anisotropy data well?
> > > How does it compare with Kriging in Arc/Info? or other geostatistics
> > > software customized to do kriging or other geostatistics functions?
> > >
> > > I tried Easykriging, a geostatistics tool developed for Matlab. It has
> > > very nice GUI, but it does not provide library which i can call in my programs.
> > > so it is good for doing it mannually for small number of data set. but if
> > > i have large number of data set, and i want to automate the kriging
> > > process for those data set, Does anyone have good recommendation on what
> > > software to use?  something that can be called in c/c++ code(or matlab)
> > > and freeware:) is preferred..
> > >
> > > Thanks a lot in advance for any input.
> > > yan
> > >
> > > ______________________________________________
> > > R-help at stat.math.ethz.ch mailing list
> > > http://www.stat.math.ethz.ch/mailman/listinfo/r-help
> >