MARK's POPAN chase.inp example 12.4.1 not working?

[geofgivens]

I have been fiddling with some POPAN modeling in RMark, and my code
was working several weeks ago. Since then, I updated to the newest
MARK version, reran my code, and it didn't work. In the process
of diagnosing my problem, I decided to use the plain MARK program
interface and walked through the POPAN example (section 12.4.1 in the
book) using the chase.inp example data. Replicating that example, I
get an error that not enough time interval values are
specified. However, the output message also shows that 7 intervals were
specified, which should be correct for the 8 occasions in this
example, and matches the book. This same error message (not enough
intervals) is what I am now encountering in RMark for my own code,
which spurred this investigation.

I'm enclosing the output from the book's example. Any ideas what is
going on? Thanks.

Geof


Program MARK - Survival Rate Estimation with Capture-Recapture Data
gfortran(Win64) Vers. 8.1 Mar 2016 12-Apr-2016 11:43:08 Page 001
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

This version was compiled by GCC version 5.3.0 using the options:
-cpp -iprefix c:\tdm-gcc-64\gcc\bin\../lib/gcc/x86_64-w64-mingw32/5.3.0/ -D_MT
-U_REENTRANT -D IEEE -m64 -mtune=generic -march=x86-64 -mthreads -O2
-fimplicit-none -fbounds-check -funroll-loops -ftree-vectorize
-ffpe-summary=invalid,zero,overflow,underflo

This problem will use 6 of 6 possible threads.


INPUT --- proc title try3;

CPU Time in seconds for last procedure was 0.00


INPUT --- proc chmatrix occasions=8 groups=1 etype=POPAN mixtures=2
INPUT --- NoHist hist=387;

INPUT --- glabel(1)=Group 1;

INPUT --- time interval 1.5 1 1 1 1 1 1.5;

ERROR -- Not enough time interval values specified.

[cooch]

I'm late for class (or will be if I spend too long on this), but, I suspect this is because of a change Gary made some time ago, in fact, but I don't remember the version), having to do with unequal intervals are handled for certain data types. See for example, section 15.9 in Chapter 15 (the robust design chapter).

Gary will need to weigh in, here.

[geofgivens]

Running the example with equal time intervals (i.e., replacing the 1.5's with 1's) yields the same error.

Yes, section 15.9 is interesting but I don't think it resolves the current problem.

[Bill Kendall]

I get the same error. At first I thought it was due to extra blank lines at the bottom of the data set, but that didn't help. The error refers to line 53 of the data, but I didn't see anything obvious. It must be a dataset problem, because I have been running that model structure frequently.

[geofgivens]

I don't know about that hypothesis. I tried it again with the chase89.inp supplied with the program, and got the same error. (My original attempt was with the chase.inp supplied with the book).

Also, I was motivated to post because I had RMark code that was working, and suddenly stopped working after I upgraded MARK. That was a POPAN model, but a totally different dataset (actually 100 simulated datasets). FYI, in those analyses there were unequal time intervals.

I no longer have my previous installation (which was several years old; I don't know which version) so I can't run the same analysis in both versions.

I'm not a MARK expert so I'm somewhat stymied about what to do.

[cooch]

I don't know about that hypothesis. I tried it again with the chase89.inp supplied with the program, and got the same error. (My original attempt was with the chase.inp supplied with the book).

Also, I was motivated to post because I had RMark code that was working, and suddenly stopped working after I upgraded MARK. That was a POPAN model, but a totally different dataset (actually 100 simulated datasets). FYI, in those analyses there were unequal time intervals.

I no longer have my previous installation (which was several years old; I don't know which version) so I can't run the same analysis in both versions.

I'm not a MARK expert so I'm somewhat stymied about what to do.

Try using MARK 7.2, which you can download from http://www.phidot.org/software/mark/downloads/.

Failing that, I have archive copies of most point releases of MARK between 7 -> 8.

[geofgivens]

Same error in 7.2. However, this time it explicitly mentions line 53 (as above). However, I think that's a red herring. Can someone else verify that the example is not working in MARK 7.2?



Program MARK - Survival Rate Estimation with Capture-Recapture Data
gfortran(Win64) Vers. 7.2 Apr 2014 12-Apr-2016 14:51:03 Page 001
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

This version was compiled by GCC version 4.8.1 using the options
-m64 -mtune=generic -march=x86-64 -mthreads -O2 -fimplicit-none -fbounds-check -funroll-loops -ftree-vectorize -fopenmp.
This problem will use 6 of 6 possible threads.


INPUT --- proc title ex7.2;

CPU Time in seconds for last procedure was 0.00


INPUT --- proc chmatrix occasions=8 groups=1 etype=POPAN mixtures=2
INPUT --- NoHist hist=387;

INPUT --- glabel(1)=Group 1;

INPUT --- time interval 1.5 1 1 1 1 1 1.5;
Error around encounter history number 53

ERROR -- Encounter history was too short for the number of occasions specified.

[jlaake]

Following works for me but I have a 2 year old version of MARK

Code: Select all

library(RMark)
chase=convert.inp("chase.inp",group.df=data.frame(age=c("adults","jacks")))
mark(chase,model="POPAN",groups="age")

Code: Select all

   Program  MARK  - Survival Rate Estimation with Capture-Recapture Data
   gfortran(Win32) Vers. 8.0 Jun 2014   12-Apr-2016 14:16:01    Page  001
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 
   This version was compiled by GCC version 4.9.2 using the options
   -m32 -mtune=generic -march=x86-64 -mthreads -O2 -fimplicit-none -fbounds-check -funroll-loops -ftree-vectorize -fopenmp.
   This problem will use 7 of 8 possible threads.

[geofgivens]

Unfortunately I get

Code: Select all

> chase=convert.inp("chase.inp",group.df=data.frame(age=c("adults","jacks")))
> mark(chase,model="POPAN",groups="age")
STOP ERROR
MARK did not run properly.  If error message was not shown, re-run MARK with invisible=FALSE

Model output is not available

Error in `colnames<-`(`*tmp*`, value = 1:7) :
  attempt to set 'colnames' on an object with less than two dimensions


This is with the current version of MARK. The error msg matches the one I got when I was originally spurred to write this query. If you split up the analysis process in RMark so you use the make.mark.model() function separately, what happens is that it displays the MARK msg about time intervals. That's what drove me to try the MARK example. (I infer that the colnames complaint occurs because make.mark.model() failed.) But to be clear, if there is a problem (rather than me just doing something dumb), then it appears to be with MARK, not RMark.

Jeff, are you running v8.0?

[jlaake]

See my last post. It is from June 2014 and says 8.0.

If you are using RMark, make sure there are no old mark32.exe, mark64.exe or mark.exe laying around that are being picked up in the path.