www.phidot.org

Hello,

I'm trying to import an .Rinp file into MARK to calculate c-hats for a multi-state model I ran in RMark, similar to these past posts:
viewtopic.php?f=2&t=3629
viewtopic.php?f=2&t=3837

Every time I attempt to import my multi-state model, MARK crashes. Unlike the other past posts, I don't see any error messages, MARK just closes after a few minutes. Please let me know if there is a way to access the error messages. MARK does create a DBF, FPT, and CDX file in the same desktop folder when I attempt to import the .Rinp file. If I try to open the DPF file in MARK I get this error message:
Error Code: 50 []
Subsystem: VO-CODE
Error Subcode:5333
Function:
Argument NUmber: 3
CallStack:
MYTOPAPPWIN:FILEOPEN (Line:109)
MYTOPAPPWIN:OPENFILE (Line:3)
APP:START (LINE:1141)
START (Line:6)

I have followed the recommendations in those posts. My files are on a folder on the desktop on my personal computer where I have access permissions. I ran the examples with the dipper, mallard, and mstrata data and they all worked normally. I delete all the files in the desktop folder every time.

My multi-state model has a large number of parameters - I have 44 weekly occasions, three states, two groups, four seasonal temporal constraints (11 weeks per season), and some parameters are fixed. I am on a Windows 10 machine with 16 gb ram using MARK version 6.2 Build 9200, RStudio Version 1.4.1106, R version 4.0.2, RMark 2.2.7.



Please let me know if there is any more information I can provide to help you help me!

Is it possible for you to try on a different computer to see if that is the problem. At least one person found that to be the case when I reviewed all of the messages relevant to this subject. If that fails this is something that Gary will have to address as I have no way of debugging it. If you want, send me (jefflaake@gmail.com) the .Rinp, .inp and the output file and I'll look at them to see if I see anything obvious.

--jeff

I worked this out with Monica off list. The problem was the number of fixed parameters. If you are setting lots of Psi parameters to 0, use the argument mlogit0=TRUE in the call to mark or mark.wrapper. This will reduce the number of real parameters and the number of fixed parameters. Also, your models will run much more quickly.