[ADMB Users] how could I get the sample size in the ADMB output files?
Yuying Zhang
yzhang2 at umassd.edu
Thu Mar 10 11:07:27 PST 2011
Hello,
I am running the ASAP (age structured assessment program) model, which
is written by ADMB.
I did the simulation twice with slightly different input .dat files.
I need the mean values, std values and sample sizes to conduct t tests
(to see if the parameters in 2 runs are significant different or not).
After basic runs (using MLE), the mean values and std values of
parameters could be found in the .std file. However, the sample sizes
for these parameters haven't been listed.
Would someone please to tell me where I can get/derive the sample sizes
from the output files of basic runs?
##################################################
I read something about the Hessian matrix.
If there is no way to get the sample size in the output files, my
solution is to count the input data. Is my guess correct?
For example, if we have a 25-year 4-fleet landing data, we also have
their age distributions (8 age bins, probabilistic -proportional
distributions p0, p1,...p8, sum(pi)=1), we also have 3 surveys for 17
years.
Is the sample size:
25*4+25*4*8+3*17=951?
or
25*4+25*4*(8-1)+3*17=851?
or
25*4+25*4+3*17=251?
###################################################
If I can't get the sample size from the basic runs, the last solution I
can think is to use MCMC. But I wish there is a way to get sample size
from basic run.
#################################################
Many thanks.
Yuying Zhang
More information about the Users
mailing list