[R-sig-Geo] pois.binom error with count data (geoRglm)

Karen Updegraff kupdegra at d.umn.edu
Mon Oct 1 21:59:14 CEST 2012


When attempting binom.krige() on a set of count data with the following call:

bktune<-binom.krige(geodata=must51_geo,
    mcmc.input=list(S.scale=0.34,n.iter=5000,thin=1),
    krige=plmodel)
#
#I get this output:
"
locations need to be specified for prediction; prediction not performed
Error in chol.default(solve(invcov + diag(fisher.l))) :
  the leading minor of order 53 is not positive definite
"

I have no clue where this error is coming from. Possibly related to
zero-loading?
pois.krige() on the same dataset, with the same model specs, runs
without errors.
Do the model parameters need to be different? Data need to be transformed?
.................
Some inputs:
-----------
plmodel (mostly default settings):
$type.krige [1] "ok"
$beta [1] 1
$cov.model [1] "matern"
$cov.pars [1] 1 1
$kappa [1] 0.5
$lambda [1] 0
attr(,"class") [1] "krige.geoRglm"

The count data:
--------------
must51_geo$data: #no overlapping points
  [1]  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
 [26]  0  0  0  0  0  1  0  0  0  1  0  1  0  0  0  0  0  0  0  0  1  0  0  0  0
 [51]  0  0  4  0  0  0  0  0  0  0  0  0  1  0  0  0  4  0  0  0  0  0  0  0  2
 [76]  1 14  2  0  0  0  0  0  0  0  0  0  0  0  0  4  0  0  0  0  0  0  0  0  0
[101]  0  0  0  0  0  2  1  2  1  0  0  1  0  0  0  0  0  0  0  0  1 32  4  0  1
[126]  1  0  1  0  0  0  0  0  0  0  0  0  0  0  2  3  2  0  0  0  0  0  0  0  0
[151]  0  0  0  1  2  0  1  0  0  0  0  0  0  0  0  0  0  1  0  0  0  5  0  0  0
[176]  0  0  0  0  0  0  0  0  0  0  0  0  0  1  0  0  0  0  0  0  0  0  0  0  0
[201]  0  0  2  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
[226]  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  1  0  0
[251]  0  0  0  0  0  0  0  0  0  0  0  0  0  1  0  0  0  0  0  0  0  0  0  0  0
[276]  0  0  0  0  0  0  0  0  1  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0


Would appreciate any suggestions.
-- 
Karen



More information about the R-sig-Geo mailing list