Problems with number of parameters in some models

questions concerning analysis/theory using program MARK

Problems with number of parameters in some models

Postby Pepe GR » Mon Jun 09, 2008 4:46 am

Hi

I was trying to run a post-fledging survival model based only in recaptures.

I have data from 15 years and I constructed (by using PIM) a model where survival was time dependent for nestlings and constant for the other ages -first diagonal values ranging 1 to 14 and value 15 for other cells- and time dependence in recapture probabilities (values 16 to 29). Thus, I had a model with 29 parameters.

My problem arose when I tried to construct the same model by using the design matrix. I constructed the model in the same way I did in previous studies with smaller datasets (results from PIM and design models where identical in those cases) and, although model had 29 parameters, once I run it, it appeared in the browser window as only containing 28 parameters. Thus, design model had same deviance that PIM model but lower AIC because of having 1 parameter less. When I looked real parameter and beta estimates there were 29 values calculates (real parameter values closely similar to those from PIM). Where is my failure?

I have found something similar when runing a Barker model based in 36 encounter histories (and 36 intervals, of course). I created a model with time dependence for all basic parameters and although 248 parameters where estimated, only 218 were reported in the browser in the column "number of parameters" (No. Par.) Is this my fault? Did I forget something? There are estimated values for 248 parameters (real and beta...)

Finally and related with this last model (but not important for analyses) was that Parameter Index Matrix was too large -for all S, p, R, R', F, F'- and only a part of each one appeared in the screen. There is no problem for moving from side to side since there is a bottom bar that allows for lateral movement, but I have not found a side bar allowing for moving up and down and I have not could see some values of matrices which remained in the depth hehe. It is possible for this also being my fault... I'm not sure.

Thank you in advance.

Pepe
Pepe GR
 
Posts: 1
Joined: Fri Jun 06, 2008 3:35 am

Return to analysis help

Who is online

Users browsing this forum: No registered users and 2 guests