Royle Biometrics model restricted to Poisson prior?

[pcutter]

Hi,

Is the Royle biometrics model personality in PRESENCE currently limited to the poisson prior distribution assumption?

Any plans to include the negative binomial?

Thanks,
Pete

[cooch]

Hi,

Is the Royle biometrics model personality in PRESENCE currently limited to the poisson prior distribution assumption?

Any plans to include the negative binomial?

Thanks,
Pete

Both distributions are in MARK.

[pcutter]

Thanks,

I have been trying to get these working but they seem to fail with double digit count data. Is there a way to get double digit counts into these?

I've inserted a snippet from my input file below. It has labels, the count histories, four occasion-specific covariates and three individual covariates. Is there a way to tweak this to use double digits?

Code: Select all

/*ESU_116*/00.0000   1   20.19   .   132.47203   171.9416   233.605   158.5592   7342.8364   33690.7422;
/*ESU_030*/0018..   1   1.24785   72.55267   .   .   127.022   197.8844   11933.5449   73433.4688;
/*ESU_084*/..00.   1   .   .   51.13419   .   196.631   189.6386   4030.3015   28810.9824;
/*ESU_133*/..00.   1   .   .   0.05662   .   139.802   66.3876   5589.7939   18908.8184;
/*ESU_83*/..0000   1   .   .   0.84721   104.39113   196.253   363.955   14369.4658   28806.3672;
/*ESU_001*/...00   1   .   .   .   2.63755   88.01   313.3685   9167.7627   23568.1191;

[gwhite]

Peter:
MARK is expecting double-digit counts. So, for 3 occasions, if you have counts of 3, 0, and 25, the encounter history looks like:
030025 1;

Go read the MARK help file, "Occupancy Royle Counts" for ful details and a bigger example.

Gary

[throop]

I have also seen papers including a zero-inflated Poisson distribution. Are there plans to include this type of distribution?

Also, for the Royle Biometrics model in Presence, can lambda be calculated for years separately? My input file included 2 years of data, but I would like to calculate lambda uniquely for each.

And, finally, has anyone else had problems with opening the new integrated habitat suitability modeling? Presence crashes when I try to open the design matrix.

Thanks

[jhines]

I have also seen papers including a zero-inflated Poisson distribution. Are there plans to include this type of distribution?

>> no plans at the moment... with time and a little research, it could be done.

Also, for the Royle Biometrics model in Presence, can lambda be calculated for years separately? My input file included 2 years of data, but I would like to calculate lambda uniquely for each.

>> You should be able to add a year covariate and use that to compute p,lambda by year.

And, finally, has anyone else had problems with opening the new integrated habitat suitability modeling? Presence crashes when I try to open the design matrix.

>> This model is pretty new, so I suspect not many have tried it yet. I've only tried simulated data with it, so if you send me your data (pao, ps2 files) off-list, I might be able to see what's going wrong.

Cheers,

Jim