Explore glmmTMB as (optional?) backend

Open jacob-long opened this issue 6 years ago • 1 comments

I believe it may be able to model autocorrelated residuals within units, which lme4 cannot do and nlme can only do for linear models.

Feb 12 '19 20:02 jacob-long

So far, just getting convergence and non-positive-definite Hessian errors when I try to use AR1 correlation structures.

Feb 12 '19 23:02 jacob-long