[R-sig-Geo] determining values for nmax, nmin and maxdist
Tim Peterson
timjp at unimelb.edu.au
Sat Jul 2 02:34:17 CEST 2016
Hi Daniel,
I've never had much success with rules of thumb (eg variogram range). I
suggest you look into more objective approaches - eg see paper below.
When I've played with global methods similar to Abedini I've found that
cross-validation errors (with and without replacement) show a search
distance that gives a minimum error. Re local search distance
estimation (see Abedini), I suspect that for any moderate to large
problem the computations becomes prohibitive.
M.J. Abedini, M. Nasseri, D.H. Burn, The use of a genetic
algorithm-based search strategy in geostatistics: application to a set
of anisotropic piezometric head data, Computers & Geosciences, Volume
41, April 2012, Pages 136-146, ISSN 0098-3004,
http://dx.doi.org/10.1016/j.cageo.2011.08.024
Cheers
Tim
On 07/02/2016 07:47 AM, Dan Turenne wrote:
> Hello listers,
>
>
> As part of the research I am doing I'm attempting to fit some kriging models, and I was wondering if there are any "rule of thumb" guidelines for selecting the values for the nmax, nmin and maxdist parameters? Intuitively I would set maxdist equal to the range of the variogram but I'm not sure about the other parameters. Any advice on this subject would be appreciated.
>
>
> Thanks,
>
>
> Daniel Turenne
>
> University of Manitoba
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
--
--------------------
Dr. Tim Peterson
The Department of Infrastructure Engineering
The University of Melbourne, 3010 Australia
T: +61 3 8344 9950 <tel:%2B61%203%208344%209950>, M: +61 0438 385 937
<tel:%2B61%200438%20385%20937>
Dept. profile :
http://www.ie.unimelb.edu.au/people/staff.php?person_ID=141135
Research Gate : https://www.researchgate.net/profile/Tim_Peterson7
Google Scholar:
http://scholar.google.com.au/citations?user=kkYJLF4AAAAJ&hl=en&oi=ao
[[alternative HTML version deleted]]
More information about the R-sig-Geo
mailing list