Page 1 of 1

Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Mon Nov 03, 2014 9:49 pm
by cpenaherrera
Dear all,

I'm using Mark 8.0 on a Windows server 2008, and when ever I start mark and run any model it goes flawlessly. The problem arises when I retrieve any model, and then opening the Parameter index Chart it will change all the names of Nbar and Nstar for f0bar and f0star. After that I can't run any model have to exit Mark and restart it, and then instead retrieving any model I have to start build it again manually. I have no idea why this is happening, but I guess it will be good to post if anybody else is having the same problem. Anyway to fix it? I'm contacting IT to reinstall Mark8.0, but perhaps will be better to go back to 7.x?

Cheers,

Cesar Penaherrera
University of Tasmania

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Mon Nov 03, 2014 10:18 pm
by cooch
cpenaherrera wrote:Dear all,

I'm using Mark 8.0 on a Windows server 2008, and when ever I start mark and run any model ...


For what data type? What sort of analysis are you trying to run? 'Nbar' and 'Nstar' aren't particularly informative.

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Mon Nov 03, 2014 10:56 pm
by cpenaherrera
Dear Evan,

Sorry for the incomplete information. The model I'm running is an IELNE, Inmigration-emigration Logit Normal estimator, from the mark-resight section. The data I'm using is like the example on the following thread:

http://www.phidot.org/forum/viewtopic.php?f=1&t=2845

PS: The example on that thread is using Rmark, but I came back to Mark as suggested in the thread, and I'm using the same type of data.

I'm unsure why this is happening, but it is not the first time it occurs to me. It used to happen with an older version of Mark 7.x, but with one update it was corrected and never happened again until now.

Hope this information now helps :)

Cheers,

Cesar Penaherrera
University of Tasmania

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Mon Nov 03, 2014 11:08 pm
by cooch
cpenaherrera wrote:Dear Evan,

Sorry for the incomplete information. The model I'm running is an IELNE, Inmigration-emigration Logit Normal estimator, from the mark-resight section.


Actually, I knew that, but for the benefit of others...thanks for specifying.

The data I'm using is like the example on the following thread:

http://www.phidot.org/forum/viewtopic.php?f=1&t=2845

PS: The example on that thread is using Rmark, but I came back to Mark as suggested in the thread, and I'm using the same type of data.

I'm unsure why this is happening, but it is not the first time it occurs to me. It used to happen with an older version of Mark 7.x, but with one update it was corrected and never happened again until now.

Hope this information now helps :)


The only thing that comes to mind is that the problem *might* relate to a change Gary made this summer, wherein the N parameter in all of the closed captures data types and robust design derivatives was been changed to be labeled f0 to prevent users (people like me, and Larissa Bailey) from mistakenly thinking setting the N parameters equal is evaluating this hypothesis. N is still provided as a derived parameter.

If there is a problem, my guess is that some of the mark-resight routines weren't updated to reflect this change. Just a hunch. Gary will have to confirm at this point...

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Mon Nov 03, 2014 11:15 pm
by cpenaherrera
Got it... thanks for your soon reply. I guess is only time to wait that he could find and patch the glitch, if there is a any. I have requested the re-installation of the software just in case. I´ll post any finding, if any.

Cheers,

C.

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Mon Nov 03, 2014 11:40 pm
by gwhite
Yep -- my mistake from the update of N to f0. I've fixed the bug and will get an update on the web in a few days.

Gary

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Tue Nov 04, 2014 8:17 pm
by cooch
gwhite wrote:Yep -- my mistake from the update of N to f0. I've fixed the bug and will get an update on the web in a few days.

Gary


Newly patched version now online.

http://www.phidot.org/software/mark/downloads/

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Fri Nov 07, 2014 12:18 am
by cpenaherrera
Dear Gary and Evan,

Today I got again an error in mark 8.0 that seems to me like this one. I was trying to use the option "Replace data and Rerun models" from the main File menu, but it failed every time I tried to use it. Looking for the error log I found that its taking again Nbar as f0bar, so it says PIM was not specified correctly . The error log I got was this:

Program MARK - Survival Rate Estimation with Capture-Recapture Data
gfortran(Win32) Vers. 8.0 Jun 2014 7-Nov-2014 15:07:18 Page 001
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

This version was compiled by GCC version 4.8.1 using the options
-m32 -mtune=generic -march=x86-64 -mthreads -O2 -fimplicit-none -fbounds-check -funroll-loops -ftree-vectorize -fopenmp.
This problem will use 1 of 1 possible threads.


INPUT --- proc title 2013 IELNE hammherhead sharks;

CPU Time in seconds for last procedure was 0.00


INPUT --- proc chmatrix occasions=15 groups=1 etype=IELogitNormalMR
INPUT --- hist=300 NoHist;

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

INPUT --- time interval 0 0 1 0 0 1 0 0 1 0 0 1 0 0;

* * WARNING * * At least a pair of the encounter histories are duplicates.

Number of unique encounter histories read was 18.

Number of individual covariates read was 0.
Time interval lengths vary and/or not equal to 1.

Data type number is 120
Data type is Immigration-Emigration Logit-Normal Mark Resight

CPU Time in seconds for last procedure was 0.00

Program MARK - Survival Rate Estimation with Capture-Recapture Data
gfortran(Win32) Vers. 8.0 Jun 2014 7-Nov-2014 15:07:18 Page 002
2013 IELNE hammherhead sharks
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



INPUT --- /* End Input File */
INPUT --- proc estimate link=Logit varest=2ndPart;


INPUT --- model={p(.) sigma(.) nbar(s) alpha(.) nstar(.) none inipar};

INPUT --- group=1 p Session 1 rows=1 cols=3 Square Constant=1;

INPUT --- group=1 sigma Session 1 rows=1 cols=1 Square Constant=2;
DATTYP = 120
group = 1 f0bar Session 1 rows = 1 cols = 1 Square Constant = 3

Parameter specified was not one of the recognized parameters:
a', a'', alpha, c, d, delta, epsilon, epsilonA, epsilonAB, epsilonB, epsilonBA,
f, F, F', f0, gamma, gamma', gamma'', gammaA, gammaAB, gammaB, gammaBA, lambda,
M, N, Nbar, Nstar, omega, p, p1, p2, pA, pB, pent, phi, phi0, pi,
psi, psi1, psi2, psiA, psiAB, psiB, psiBA, psiBa, ptilde, R, R', r, rA,
rAB, rAb, raB, rBA, rBa, S, sigma, sigmaf, sigmap, sigmaphi, sigmar, sigmaS,
theta, theta', U, or VarAdd


INPUT --- group=1 f0bar Session 1 rows=1 cols=1 Square Constant=3;

ERROR -- Type of Parameter Input Matrix (PIM) was not specified correctly.


Hope it helps...

Cheers,

Cesar Penaherrera
University of Tasmania

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Fri Nov 07, 2014 8:46 pm
by gwhite
I realized during the 2-hour upload from my motor home that I've forgotten to make the change in the rerun models routine. I had hoped that nobody would discover this problem until I got home tonight and a faster connection. Anyway, a new version is on the web that fixes this additional bug.

Garu

Re: Mark changes the Nbar and Nstar for f0bar and f0star

PostPosted: Sun Nov 09, 2014 11:51 pm
by cpenaherrera
Many many thanks Gary for all your hard work on this! :)