"fix derived parm. output in multi-season models"

announcements (new versions, changes, bugs, installation problems...) related to program PRESENCE

"fix derived parm. output in multi-season models"

Postby nathanvs » Fri Nov 25, 2016 8:38 pm

Latest update, v11.6, lists that it "fix[ed] derived parm. output in multi-season models." I had been using some of these derived parameter outputs in some results I had drafted up--what exactly was the bug/issue with the output that was fixed? Wondering if I will need to re-run these now.

Thanks!
nathanvs
 
Posts: 6
Joined: Fri Nov 20, 2015 5:07 pm

Re: "fix derived parm. output in multi-season models"

Postby jhines » Sat Nov 26, 2016 10:35 am

There was no problem with derived parameters in earlier versions. At some point, I changed the code in the multiseason model and it stopped printing derived parameters. The most recent fix just put those estimates back in the output.
jhines
 
Posts: 386
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: "fix derived parm. output in multi-season models"

Postby nathanvs » Mon Nov 28, 2016 6:18 pm

Thank you, great to know. I am also having some trouble in 11.6 with the model averaged estimates of derived parameters for the simple multi-season model. The normal parameters are coming through fine, but there are several thousand extra sites being estimated for the model-averaged parameter estimates, most of which are empty. Some of these extra sites are labeled with their "site name" as the other parameters; others have no site name. Many of the actual sites are missing the derived parameter output as well. I assume this is some sort of bug?

For example (with breaks to skip around in the file):

psi( 2)(site 1) model-avg-est: . se= . 15
psi( 2)(site 2) model-avg-est: . se= . 18
psi( 2)(site 3) model-avg-est: . se= . 21
psi( 2)(site 4) model-avg-est: . se= . 56
psi( 2)(site 5) model-avg-est: . se= . 110
.......
psi( 2)(site 113) model-avg-est: . se= . 297a-12
psi( 2)(site 114) model-avg-est: . se= . 297b-12
psi( 2)(site 115) model-avg-est: 0.6018 se= 0.0715 298-06
psi( 2)(site 116) model-avg-est: . se= . 299-06
psi( 2)(site 117) model-avg-est: . se= . 300-06
.......
psi( 2)(site 235) model-avg-est: 0.1351 se= 0.0426 Y-33
psi( 2)(site 236) model-avg-est: 0.2657 se= 0.0515 Y-33a
psi( 2)(site 237) model-avg-est: 0.2746 se= 0.0519 Y-33b
psi( 2)(site 238) model-avg-est: 0.3961 se= 0.0712 Y-34
psi( 2)(site 239) model-avg-est: 0.2895 se= 0.0535 Y-34a <-- this is the last real site
psi( 2)(site 240) model-avg-est: . se= .
psi( 2)(site 241) model-avg-est: . se= .
psi( 2)(site 242) model-avg-est: . se= .
psi( 2)(site 243) model-avg-est: . se= .
psi( 2)(site 244) model-avg-est: . se= . psi1
psi( 2)(site 245) model-avg-est: . se= .
psi( 2)(site 246) model-avg-est: . se= .
psi( 2)(site 247) model-avg-est: . se= .
psi( 2)(site 248) model-avg-est: . se= . psi1
psi( 2)(site 249) model-avg-est: . se= . gam(1)
psi( 2)(site 250) model-avg-est: . se= . gam(2)
.......
psi( 2)(site 662) model-avg-est: . se= .
psi( 2)(site 663) model-avg-est: . se= .
psi( 2)(site 664) model-avg-est: . se= .
psi( 2)(site 665) model-avg-est: 0.1841 se= 2.8687
psi( 2)(site 666) model-avg-est: 0.3672 se= 3.8087
psi( 2)(site 667) model-avg-est: 0.5423 se= 4.8386

As you can see, the standard errors for the "blank" sites are also turning up very large.
nathanvs
 
Posts: 6
Joined: Fri Nov 20, 2015 5:07 pm

Re: "fix derived parm. output in multi-season models"

Postby jhines » Tue Nov 29, 2016 8:42 am

Definitely looks like a bug in the model-averaging routine. Can you send me the most recent backup zipfile from your project folder? Thanks,

Jim jhines at usgs dot gov
jhines
 
Posts: 386
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: "fix derived parm. output in multi-season models"

Postby jhines » Thu Dec 01, 2016 2:06 pm

I've updated PRESENCE to ver. 11.7, which fixes the model-avg. derived parameter output.
jhines
 
Posts: 386
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA


Return to software problems/news

Who is online

Users browsing this forum: No registered users and 1 guest