[R] Function for linear mixed model with gamma-distributed random effects?

[Dieter Menne]

Christopher Hatherly <Christopher.Hatherly <at> anu.edu.au> writes:

I am trying to fit an ellipse equation to some spatial human factors data, 
varying the major and minor axes  randomly and specifying an exp~ variogram 
for errors. Using normally-distributed random effects produces some -ve
minor/major axes. I am hoping to be able to specify a gamma distribution to
overcome this, but have not been able to work out how.

Does anybody know an R function that implements lme with gamma-distributed
random effects?
---------------

One of the function in DPpackage might give you a starter. Or use RBrugs, but
that will required some additional reading. Or better, try to fit you model in
some existing statistical context, what you want to do it seem a bit hand-knit
to me and resolvable via book-lookup.

Dieter