I am having difficulty understanding the output from a likfit call,
specifically the output for the nugget. When the partial sill is non-zero,
the estimated nugget that is returned is zero. When the partial sill is zero,
I get a non-zero nugget. The following output may be helpful:
Estimation method: maximum likelihood
Parameters of the mean component (trend):
beta0 beta1 beta2 beta3 beta4 beta5
2.4299 2.5095 4.8184 -0.0084 -0.0625 -0.0057
Parameters of the spatial component:
correlation function: spherical
(estimated) variance parameter sigmasq (partial sill) = 1694
(estimated) cor. fct. parameter phi (range parameter) = 32.1
anisotropy parameters:
(fixed) anisotropy angle = 0 ( 0 degrees )
(fixed) anisotropy ratio = 1
Parameter of the error component:
(estimated) nugget = 0
Transformation parameter:
(fixed) Box-Cox parameter = 1 (no transformation)
Maximised Likelihood:
log.L n.params AIC BIC
"-98.92" "9" "215.8" "224.8"
non spatial model:
log.L n.params AIC BIC
"-101.5" "8" "219.0" "226.9"
Call:
likfit(geodata = geodataK, trend = "2nd", ini.cov.pars = c(1700,
50), cov.model = "sph", method.lik = "ML")
---
This is the code I used:
geodataK <- as.geodata(data[,c(2,3,11)], coords.col=1:2, data.col=3)
geodataK
bin4 <- variog(geodataK, uvec=seq(0,163.44,l=21), max.dist=50,
estimator.type="modulus", trend="2nd"); plot(bin4, main =
"(f) Potassium",
xlab = "", ylab = "")
mod1 <- likfit(cov.model="sph",geodataK,
trend="2nd",ini=c(1700,50),
method="ML");summary(mod1)
lines(mod1, lty=1)
---
I am also trying to figure out how to calculate R^2 values for the likfit
models that I fit to the semivariogram.
I am using R version 1.9.1 (rw1091) and geoR version 1.4-8 on a PC running MS
Windows XP Professional version 2002.
Many thanks,
Melanie Link-Perez
Miami University
Paulo Justiniano Ribeiro Jr
2004-Sep-29 01:32 UTC
[R] not understanding geoR "nugget" output
Melaine When estimated phi=0 or sigmasq=0 the odel is a "pure nugget effect" and you cannot distinguish between sigmasq and tausq Therefore it is a convention in geoR to assign the estimated varioance to tausq. Regarding R^2: forgaussian models you can compute this values using the maximised likelihood and othe model information Alternatively you can use likfit with the argument components=T. This will return the estimated model components frwom which you can compute R^2. Since this is a package specific question feel free to contact me directly if you have any further queries best P.J. On Tue, 28 Sep 2004, Melanie A. Link-Perez wrote:> I am having difficulty understanding the output from a likfit call, > specifically the output for the nugget. When the partial sill is non-zero, > the estimated nugget that is returned is zero. When the partial sill is zero, > I get a non-zero nugget. The following output may be helpful: > > Estimation method: maximum likelihood > > Parameters of the mean component (trend): > beta0 beta1 beta2 beta3 beta4 beta5 > 2.4299 2.5095 4.8184 -0.0084 -0.0625 -0.0057 > > Parameters of the spatial component: > correlation function: spherical > (estimated) variance parameter sigmasq (partial sill) = 1694 > (estimated) cor. fct. parameter phi (range parameter) = 32.1 > anisotropy parameters: > (fixed) anisotropy angle = 0 ( 0 degrees ) > (fixed) anisotropy ratio = 1 > > Parameter of the error component: > (estimated) nugget = 0 > > Transformation parameter: > (fixed) Box-Cox parameter = 1 (no transformation) > > Maximised Likelihood: > log.L n.params AIC BIC > "-98.92" "9" "215.8" "224.8" > > non spatial model: > log.L n.params AIC BIC > "-101.5" "8" "219.0" "226.9" > > Call: > likfit(geodata = geodataK, trend = "2nd", ini.cov.pars = c(1700, > 50), cov.model = "sph", method.lik = "ML") > > > --- > This is the code I used: > > geodataK <- as.geodata(data[,c(2,3,11)], coords.col=1:2, data.col=3) > geodataK > bin4 <- variog(geodataK, uvec=seq(0,163.44,l=21), max.dist=50, > estimator.type="modulus", trend="2nd"); plot(bin4, main = "(f) Potassium", > xlab = "", ylab = "") > mod1 <- likfit(cov.model="sph",geodataK, trend="2nd",ini=c(1700,50), > method="ML");summary(mod1) > lines(mod1, lty=1) > > --- > > I am also trying to figure out how to calculate R^2 values for the likfit > models that I fit to the semivariogram. > > I am using R version 1.9.1 (rw1091) and geoR version 1.4-8 on a PC running MS > Windows XP Professional version 2002. > > > Many thanks, > Melanie Link-Perez > Miami University > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html > >Paulo Justiniano Ribeiro Jr Departamento de Estat??stica Universidade Federal do Paran?? Caixa Postal 19.081 CEP 81.531-990 Curitiba, PR - Brasil Tel: (+55) 41 361 3573 Fax: (+55) 41 361 3141 e-mail: paulojus at est.ufpr.br http://www.est.ufpr.br/~paulojus /"\ \ / Campanha da fita ASCII - contra mail html X ASCII ribbon campaign - against html mail / \