R help - Aug 2011 - looking for REML from a gnls fit or a nonlinear function (without random effects) fit using REML

Dear R users,
      I am fitting nonlinear mixed effects models with autocorrelated errors
(an AR(1) model on the residuals) using NLME and am comparing a set of
models that contain the same fixed effects structure but with different or
no random effects (nested) .  The issue I've come across that I'm unable
to
fit the model with fixed effects only (i.e. no random effects) using
restricted maximum likelihood (REML) in order to appropriately compare the
models with and without random effects.  Does anyone know of a function that
will fit nonlinear models using REML or that will extract a restricted
log-likelihood from a gnls object?


Daniel K. Okamoto
PhD Student
1206 Marine Science Research Building
Department of Ecology, Evolution and Marine Biology
University of California, Santa Barbara
okamoto@lifesci.ucsb.edu

	[[alternative HTML version deleted]]