Dear Kai, you might want to look at function autofitVariogram in R package automap, which exactly tries to do this. It uses reasonable default (starting) values, but as any algorithm, it may not work well with particular input data that is extreme in some sense. -- Edzer
On 05/10/2010 06:52 AM, Kai Zhang wrote: > Hi folks, > > I am working on a problem dealing with fitting variogram automatically due to > a large dateset. Gstat R package is used to conduct this work. Part of this > work is to compare performance of different kriging methods. The comparison > criteria is RMSE. I am using fit.variogram function for ordinary kriging, and > fit.lmc for co-kriging. I have to specify a theoretical variogram at first. > The problem is that results are sensitive to initial values specified in > theoretical variograms, mainly "Range". Gstat manual points out "If fitting > the range(s) is part of the job of this function, the results may well depend > on the starting values, given in argument model." (for fit.variogram > function). My question is how to choose starting values. I did take a few > data to check variograms, and found they had different "range" values. I am > not sure what starting values I should choose, say a small "range" value or a > large "range" value. > > Best regards, > Kai > > > > > __________________________________________________ > 赶快注册雅虎超大容量免费邮箱? > > _______________________________________________ > R-sig-Geo mailing list > R-sig-Geo@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 http://www.52north.org/geostatistics e.pebe...@wwu.de _______________________________________________ R-sig-Geo mailing list R-sig-Geo@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-geo
