Spatially autocorrelated random effects in logistic regression.
Hi Hugh, Check out spGLM in the spBayes package. This should do what you need. Please let me know if you have any questions. -Andy Quoting Hugh Sturrock <hughsturrock at hotmail.com>:
Hi all, I am trying to model binomial data (infection status) of people living in households. I have run a logistic regression with a household level random effect to account for correlation between individuals within households. When I look at the value of the
random
effect (intercept) at each household there appears to be residual spatial autocorrelation (using Moran's I and semivariograms) between
households. I would like to repeat the analyses accounting for this spatial autocorrelation between households, i.e. between random effects. I'm not a mathematician? but having done some reading it seems as if this can't be done using the command glmmPQL in MASS,
but
perhaps can be done using the geoRglm package. BUT, having read through the geoRglm vignette, I'm still pretty clueless as to how to
code such a model. If anyone could give me an example I would be MASSIVELY appreciative. Thanks!Hugh ? ? ? ? [[alternative HTML version deleted]]
_______________________________________________ R-sig-Geo mailing list R-sig-Geo at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo