[R-sig-Geo] local Universal Kriging with krige()

Edzer Pebesma edzer.pebesma at uni-muenster.de
Tue Aug 26 08:41:21 CEST 2008


Hi Sven,

I assume that GEO_UNIT is a categorical variable. Without having seen 
your data, my guess is that the error arises when you run into a 
location where the local neighbourhood selection does not contain all 
the GEO_UNIT levels anymore, so it can't estimate all of the trend 
parameters from that selection. Do you have a specific reason why you 
want to  make these parameters neighbourhood dependent?

Best wishes,
--
Edzer

Sven Burbeck wrote:
> Dear all,
> I have a problem when trying to perform local Universal kriging with krige().
> The background:
> For model definition I have to use a larger dataset since this is the only
> one available with sufficient information. However, the interpolation
> should only be performed for a smaller subset and as local UK (although I
> use the global datset for model definition, I know that the data has local
> differences).
>
> MyExample:
> myResult <- krige(log(myValue) ~ GEO_UNIT - 1,
>
> locations = myGlobalDataset,
>
> newdata = myPredictionGrid_1000_utm,
>
> model = myVgmf,
>
> nmax = Inf,
>
> nmin = 0,
>
> maxdist = 18000,
>
> indicators = FALSE,
>
> na.action = na.pass,
>
> debug.level = -1)
>
> I get the following error:
> "solve.c", line 87: singular matrix in function Usolve()
> "lufactor.c", line 207: singular matrix in function m_inverse()
> Fehler in predict.gstat(g, newdata = newdata, block = block, nsim = nsim,  :
>   m_inverse
>
> When "maxdist = Inf" everything works well.
>
> Any hint where the problem could be would be welcome.
>
> Sven
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>   

-- 
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of Münster,
Weseler Straße 253, 48151 Münster, Germany.  Phone: +49 251
8333081, Fax: +49 251 8339763  http://ifgi.uni-muenster.de/




More information about the R-sig-Geo mailing list