[ADMB Users] multiple random effects vectors with correlated Gaussian priors
hskaug at gmail.com
Mon Mar 29 05:06:39 PDT 2010
This feature of ADMB is not much used and tested. The only
real model as far as I know is:
which probably is your template. The manual is written
post-hoc, and may be incorrect if you say so.
The other feature for doing spatial modelling is:
This is more recent work, and looks more frightening.
It is computationally more powerful, and gives you more
I will try to answer your questions as far as I can, which unfortunately
is not long. If you choose to pursue this type of model,
it would be nice if you share your experience, and update
the user manual as you see it needed.
> (*) Following section 4.5 in the ADMB-RE manual, I should define the
> covariance matrix inside a SEPARABLE_FUNCTION. However, the "spatial"
> example defines a NORMAL_PRIOR_FUNCTION instead, but I don't see this
> anywhere in the manual. Are NORMAL_PRIOR_FUNCTION and
> SEPARABLE_FUNCTION equivalent constructs? The spatial example also
> defines a regular FUNCTION called "evaluate_M" that calls the
> NORMAL_PRIOR_FUNCTION. This all seems fairly undocumented, so I'm
> wondering if this is the necessary way to write these types of models.
> I've googled and can't seem to find much on these models in ADMB.
The manual may be wrong, use the example.
> (*) Is it possible to specify correlated priors for multiple vectors
> of random effects?
I am not sure. Only Dave would know. It would be nice
if the answer is yes.
> (*) Is it possible to construct a concentrated likelihood for the
> fixed effects when computations are done inside of a
I take "concentrated likelihood" to mean "integrated likelihood".
Is this also for the spatial model, or is it for a separate model?
More information about the Users