Dear Richard,
I'm aware of probit/tobit spatial econometric models, but not of GLM-based
spatial-lag models.
Personally, I use Dan Griffith's spatial filtering techniques (also implemented
- by Michael Tiefelsdorf - in spdep), which allow me to use GLMs with any
underlying statistical distribution (in my case, I used logistic, Poisson and
negative binomial regressions).
Of course, this doesn't allow you to explicitly model the spatial
autocorrelation of the dependent. Still, the residuals are cleaned up, and you
used the (hopefully) right underlying statistical distribution...
Good luck!
Roberto
********************
Roberto Patuelli, Ph.D.
Post-doc researcher
Institute for Economic Research (IRE)
University of Lugano (USI)
via Maderno 24, CP 4361
CH-6904 Lugano
Switzerland
Phone: +41-(0)58-666-4166
Email: [EMAIL PROTECTED]
********************
**************************
Hi,
I am interested in estimating the spatial autoregressive lag model
y = rWy +Xb + e
where W is a square matrix of weights, r an autocorrelation parameter, b
a vector of coefficients, X a matrix of covariates.
There are several functions/packages in R that can estimate this model
(lnam in sna, lagsarlm in spdep). They work well. However, these all
assume normality and linearity.
I would like to run this model with y being a count or a probability.
This would "generalize" the autoregressive lag model. Unfortuntely, I
can't seem to find a function/package that is able to estimate this
exact model. Can anyone point me in the right direction?
thanks, Richard
BTW, I would also be interested in a package that can estimate a
generalized error model.
********************
Roberto Patuelli, Ph.D.
Post-doc researcher
Institute for Economic Research (IRE)
University of Lugano (USI)
via Maderno 24, CP 4361
CH-6904 Lugano
Switzerland
Phone: +41-(0)58-666-4166
Email: [EMAIL PROTECTED]
********************
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
