model averaging crash

questions concerning analysis/theory using program MARK

model averaging crash

Postby jsaracco » Wed Sep 13, 2006 5:34 pm

Hi. I am running open-population cjs-type models, estimating phi and p. When I try and run the model averaging platform, I get the following message: "Unique ID > 8000. Program may crash? Tell Gary." Other info that might be helpful...PIMs are set up such that I have 48 groups, 15 sampling occasions, and 216 parameter index numbers (although from the design matrix most of these are estimated the same - there are only 24 parameters estimated by my most general model). Maybe I'm doing something wrong??

Any insight would be appreciated. Thanks!

Jim
jsaracco
 
Posts: 4
Joined: Tue Mar 28, 2006 1:07 pm

model averaging error

Postby brouwerl » Tue Apr 17, 2007 3:45 am

I am running a multi state model with 6 groups and 29 occasions, when i try to get model averaged values i get the samne error: "unique id>8000, program may crash? tell Gary"
Is this because there are too many parameters? Is there a way to solve this?

Lyanne
brouwerl
 
Posts: 10
Joined: Thu Nov 04, 2004 5:22 am
Location: Trondheim, Norway

Re: model averaging error

Postby cooch » Tue Apr 17, 2007 7:02 am

brouwerl wrote:I am running a multi state model with 6 groups and 29 occasions, when i try to get model averaged values i get the samne error: "unique id>8000, program may crash? tell Gary"
Is this because there are too many parameters? Is there a way to solve this?

Lyanne


As per the MARK FAQ:

MARK isn't working - I keep getting a <insert weird gibberish code> error message - what should I do?

Well, first, don't panic. The next thing you should do is re-install MARK, making sure you're using the most recent version of the software (see above on where to get MARK). When re-installing MARK, make sure that (i) you first completely uninstall the old version (using the Windows Uninstall Programs method), and then deleting the MARK subdirectory that was created when you originally installed MARK, and then (ii) installing MARK from the most recent build. If after re-installing MARK you still get the same error messages, try a different analysis. If the error is 'analysis-specific', then it may (or may not) be related to your input file. Have a look to see if there are any problems with it. Then, if none of the above works, post your question here.
cooch
 
Posts: 1654
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Unique ID >8000

Postby abreton » Tue Apr 17, 2007 4:38 pm

I thought this error sounded familiar so I went digging through my correspondence with Gary White going back about 4 years...

Here is the question I asked him (on 8/26/2003):

"I selected Model Average under Option from the Main Menu and immediately the following message popped up: 'Unique ID >8000. Program may crash? Tell Gary'."

And here is his response:

"If you get this message when you have just opened MARK, and then select the Model Average option, then you will have to go to File | Preferences, and change the Model Averaging Preference to batch mode.
Otherwise, I suspect that you had been working a while in MARK before you selected Model Average, and so a bunch of the first 8000 Unique ID's had already been burned up. In particular, PIM windows are hard on this list. Not opening any PIMs before Model Average would be a good bet."
abreton
 
Posts: 111
Joined: Tue Apr 25, 2006 8:18 pm
Location: Insight Database Design and Consulting

Postby brouwerl » Wed Apr 18, 2007 12:35 pm

Using the non-interactive specification window indeed works, thanks!
brouwerl
 
Posts: 10
Joined: Thu Nov 04, 2004 5:22 am
Location: Trondheim, Norway


Return to analysis help

Who is online

Users browsing this forum: No registered users and 1 guest

cron