Dear members of the list,
I'm fitting ordinal regressions using polr, and in some models I
get the error copied below. Dependent variable is an ordered factor
of bird abundance categories, and predictors are continuous habitat
variables.
> ro6 <- polr(formula = abun ~ InOmbrot + Oliva.OC + ToCultAr +
DivCulArb + AltitMax + COORXY)> summary(ro6)
Re-fitting to get Hessian
Error in La.svd(x, nu, nv, method) : error code 8 from Lapack
routine dgesdd
In addition: Warning messages:
1: NaNs produced in: dlogis(x, location, scale, log)
2: NaNs produced in: dlogis(x, location, scale, log)
But if I call for the "ro6" object I get the coefficients but
without standard errors:
> ro6
Call:
polr(formula = abun ~ InOmbrot + Oliva.OC + ToCultAr + DivCulArb +
AltitMax + COORXY)
Coefficients:
InOmbrot Oliva.OC ToCultAr DivCulArb
AltitMax
-2.721242e-01 2.590153e-02 2.098157e-02 7.908437e-01 2.088895e-
03
COORXY
5.937940e-06
Intercepts:
0|1 1|2 2|3
18.16425 18.97070 20.94103
Residual Deviance: 817.667
AIC: 835.667
It seems that the error is related to the number of predictors
involved since if I use a lower number, say five predictors, polr
does not produce this error message in summary.
I have been looking at previous messages concerning errors in polr
but I haven't found the solution. I would ackowledge very much any
help.
Germ??n L??pez
