Re: [unmarked] ranef() on unmarkedFitPCO object that was fit using dynamics='trend'

110 views
Skip to first unread message

Richard Chandler

unread,
Jun 12, 2012, 9:27:51 AM6/12/12
to unma...@googlegroups.com
Looks like a bug. I'll try to release a new version later this month.

Richard

_____________________________________
Richard Chandler, post-doc
USGS Patuxent Wildlife Research Center
301-497-5696



From: Thomas Prowse <tpro...@gmail.com>
To: unma...@googlegroups.com
Date: 06/12/2012 08:44 AM
Subject: [unmarked] ranef() on unmarkedFitPCO object that was fit using dynamics='trend'
Sent by: unma...@googlegroups.com





Hi All

Is it possible to use the ranef() function on an unmarkedFitPCO object that was fit using dynamics='trend'

I get the following:

> m1 <- pcountOpen(~1, ~1, ~1, ~1, umf1, K=30, mixture = 'P', dynamics='constant')
> m2 <- pcountOpen(~1, ~1, ~1, ~1, umf1, K=30, mixture = 'P', dynamics='trend')
> re1 <- ranef(m1)
> re2 <- ranef(m2)
Error in .local(object, ...) :
 omega is not a parameter in the dynamics='trend' model


If not, is there an another option I'm unaware of?
Cheers
Thomas



Thomas Prowse

unread,
Jun 12, 2012, 9:08:11 PM6/12/12
to unma...@googlegroups.com
Thanks Richard. On an unrelated matter - for an annual species, would it be possible to fix omega at 0 (rather than 1) in pcountOpen?

Thomas

Richard Chandler

unread,
Jun 13, 2012, 8:43:52 AM6/13/12
to unma...@googlegroups.com
Hi Thomas,

That would be possible, but it's unlikely that I will have time to implement it any time soon. One option would be to fit the model in the most recent version of JAGS. Note that neither WinBUGS or OpenBUGS can fit the model for some reason, and old versions of JAGS don't work either. Here is so some code that might help: https://sites.google.com/site/unmarkedinfo/home/workshops/patux2012/r-scripts/script-dailMadsen.R?attredirects=0&d=1

Thomas Prowse

unread,
Jun 13, 2012, 10:14:21 AM6/13/12
to unma...@googlegroups.com
Thanks Richard - that code looks very useful. I'll try this with JAGS.
Reply all
Reply to author
Forward
0 new messages