Edzer Pebesma wrote:
My guess is that from constant data, the (co)variance is constant and
zero, so the covariance matrix cannot be decomposed (hence: LDLfactor
errors).
Is this a case that autokrige should catch?
I added a check in autoKrige. The output for the example below is now:
Dear all
I want to simulate a spatially-correlated random field which follows a
uniform rather than than Gaussian distribution. Does anybody know a
straight-forward way to do this?
Nick
___
R-sig-Geo mailing list
R-sig-Geo@stat.math.ethz.ch
Thank you Edzer for giving a hint about the meaning of the 'LDLfactor'
error, and thank you Paul for adding a check to the 'autoKrige'
function.
Kind regards,
Mauricio
--
Ph.D. Candidate,
University of Trento
Dept. of Civil and Env. Engineering
Hi Nick,
Very interesting problem. At first thought, I imagined that you just want to
simulate noise ;)
In the geoR package (http://leg.ufpr.br/geoR/) there is a function to simulate
Gaussian Random
Fields (uses actually RandomFields package) using various models e.g.:
library(geoR)
Hi Paul, Edzer,
I understand why the singular matrix problem happens and I know that there is
not really a
mathematical solution around it:
x - matrix(runif(100), nrow=10)
x.i - solve(x)
str(x.i)
num [1:10, 1:10] 0.8191 -1.0293 0.0826 1.068 -0.2106 ...
# add a 'singular' column
x[,1:10] -
Hi Nick
One way is to use simulated annealing (see gslib) putting as objective
function your desired variogram and histogram.
(but I guess that by means of some data transformation you can do that
with a simple sequential gaussian simulation approach)
Bye
Sebas
At 10.06 17/11/2009, Nick Hamm
Hi Ultrich,
Facundo Muñoz apparently made a GRASS function to derive distances along
streams:
https://stat.ethz.ch/pipermail/r-sig-geo/2009-November/006851.html
17 observations only? That is really tight for any geostatistical analysis (on
top, you want to do 3
dimensions!). I would instead
Dear Paul and Edzer,
2009/11/17 Tomislav Hengl he...@spatial-analyst.net:
Hi Paul, Edzer,
I understand why the singular matrix problem happens and I know that there is
not really a
mathematical solution around it:
x - matrix(runif(100), nrow=10)
x.i - solve(x)
str(x.i)
num [1:10,
Tom, you can already do this:
library(gstat)
data(meuse)
coordinates(meuse)=~x+y
data(meuse.grid)
gridded(meuse.grid)=~x+y
meuse=meuse[c(1,1:155),] # replicate first observation
pr1 = predict(zinc~1,meuse,meuse.grid,vgm(1, Exp, 300)) # will break
# the following will generate NA's for cells where
Dear all
I have a question about the random number seeding in R.
I want to simulate several random fields. Each RF should have zero
nugget, the same sill but a different range (e.g., 100x100, range: 1
- 30). Let's stick with the Gaussian case for now. I use the
following code
I just want to point out that krige has no problem with constant
observations, as long as it is provided with a valid (i.e. semi negative
definite, or something like that) variogram model. The problem is to
automatically fit such a function from data that are constant, as they
have zero
Hi Edzer,
Thanks for the info. I was not aware that I can simply set cn_max and the
predictions will not break
(and I still do read the gstat manual).
I guess that this is then the solution to our problem.
For me it enough to generate a map with NA's, then zoom into map to see where
the
Hi,
This is not possible in the version on CRAN now, but has been changed in
the development version that you can find on one of the links below
(depending on platform), should be on CRAN soon.
http://www.intamap.org/downloads/intamap_1.3-1.zip
Hello,
I am taking the log of precipitation values and therefore many are now NA
values. I want to continue to krig my precipitation matrix.
Is there a way to ignore these values with kriging.
My attempt with is.nan still gives:
Erreur : dimensions do not match: locations 105 and data 12
I am wondering if anyone has gone through the trouble of establishing a
neighborhood analysis on a set of polygon using the tools provided by spdep
to analyse a set of points. In fact i'm still not sure that it makes sense.
(¡I want to believe it does!)
So far, I have been using the spdep
If you are interpolating precipitation you probably should not ignore
the zeros. If you want to log transform your values, perhaps you can
use log(x+1) instead of log(x). Robert
On Tue, Nov 17, 2009 at 10:43 AM, Tobin Cara cara.to...@epfl.ch wrote:
Hello,
I am taking the log of precipitation
16 matches
Mail list logo