Hi.
I'm a new user of R and I'm trying to run a spatial probit model using
the
spatialprobit package. I've a dataset of 15677 obs (municipalities followed
up from 2001 to 2012). My depvar is the decision to outsource a particular
service (binary) and my indepvar of interest is "shareholder" (equal
to 1 if
the municipality is a shareholder of a firm operating that particular
service, 0 otherwise).
I guess that there is a spatial dependence in the decision due to the
occurrence of several municipalities being shareholders of the same firm.
After having imported the weight matrix W (15677x15677, with element eij
equal to 1 if municipality i and municipality j are shareholders of the same
firm, 0 otherwise) I have tried to run the model. Here the code with the
relevant output:
>library(spatialprobit)
package 'spatialprobit' was built under R version 3.0.2
> set.seed(2)
> y<-as.numeric(X$cp)
> sprobit<-sarprobit(y ~
1+shareholder+size+municipal+ineconomia+dur1+dur2+dur3+dur4, data=X, W=W1)
Errore in .local(x, ...) :
cs_lu(A) failed: near-singular A (or out of memory)
As you can see I get this error and have tried without any result to cope
with it.
Have anyone some suggestions?
Thank you
Ric
[[alternative HTML version deleted]]