Dear Gstat users,
I have a problem while trying to cross-validate my
cokriging results:
my predictions file is like this:
data(a):'a.txt',x=1,y=2,z=3,v=6;
data(b):'b.txt',x=1,y=2,z=3,v=6; merge a with b; variogram(a): .........;
variogram(b): ..........; variogram(a,b): ...........; mask: 'xxx.asc';
set zmap= ###; predictions(a): 'aaa.asc'; It works perfectly, but when I try to
cross-validate results automatically with GSTAT with such a
program:
data(a):'a.txt',x=1,y=2,z=3,v=6;
data(b):'b.txt',x=1,y=2,z=3,v=6; merge a with b; variogram(a):
..........;
variogram(b): ..........; variogram(a,b): ............; set xvalid=1;
set output = 'a2.txt'; the program returns me the following error
message:
"chfactor.c", line 130: singular matrix in
function LDLfactor()
where is the error? Can you send me the text of a
command file for cross-validation that worked?
further information:
- I'm working on a win'98 platform.
- I tried to see in the mailing-list archive. If I
put the command set output = 'xvalid.out'; GSTAT returns a xvalid.out
file with usual predictions and variances , and the same happens with the
command line gstat -o xvalid.out xvalid.txt (my command
file).
Thank you very much for your help.
Daniele Iannuzzo
|