[ADMB Users] mceval output inconsistent with mcmc display

Ian Taylor Ian.Taylor at noaa.gov
Thu Jun 9 13:01:18 PDT 2011


Hi Alex,

To further clarify where the problem lies, you might try to read the binary
files .psv files that are produced by the -mcmc command.

 

There's a link to some example R code in
http://admb-project.org/documentation/faq#how-to-run-mcmc (page 16 of the
linked document from Anders shows application of the readBin function).

-Ian

 

 

From: users-bounces at admb-project.org [mailto:users-bounces at admb-project.org]
On Behalf Of Campbell, Alex
Sent: Thursday, June 09, 2011 1:30 AM
To: users at admb-project.org
Subject: [ADMB Users] mceval output inconsistent with mcmc display

 

mcmc stdout displays the current value of the (negative log) likelihood
(twice for some reason) then "mcmc sim x acceptance rate y1 y2" like this:

757.732553269085 757.732553269085 
 mcmc sim 401  acceptance rate 0.0698254364089776 0.14 

I've checked the code to make sure it is actually the likelihood being
displayed and it seems to be but tell me if I'm wrong. when I use mceval the
parameter values are inconsistent with this displayed likelihood (their
likelihood is much lower). How could this be happening? by the end of a
million iteration run the displayed likelihood was 603 which seems within
reason. the corresponding parameter values obtained from mceval were around
-700! and needless to say a really terrible fit. 

any idea what could be going wrong? Am I interpreting the displayed output
correctly? The current likelihood should be the current *accepted*
likelihood, so should translate to valid parameters in the saved to psv and
read during mceval.? Code and input files are rather big - I won't spam
people with it. I can send directly if anyone volunteers. This is what I'm
doing for the mceval section if that sheds any light (I doubt it). 

  if (mceval_phase()) { 
    cout << a_sr << " " << b_sr << " " << c_sr << " " << iota << " " <<
vartheta << " " << vartheta_ce << " " << nexpp << " ";

    for (int i=1;i<=nyears;i++) { 
      cout << xi(i) << " "; 
    }    
    cout << log_sigma_xi << " "; 
    for (int i=1;i<=(nyears_spatial*ncells);i++) { 
      cout << psi(i) << " "; 
    } 
    cout << log_gamma << " " << rho_psi << " "; 
    cout << log_sigma_psi << " "; 
    cout << sigma_C << " " << sigma_E << " "; 
    cout << early_kappa << " "; 
    //cout << rho_psi_cl << " "; 
    //cout << log_sigma_psi_cl << " "; 
    cout << E1980 << " " << E1985 << " " << nexpp2 << endl; 
    //cout << mu_psi_cl << " "; 
  } 

 

This has occurred both using admb 10.0 on a windoze box and admb 9.1 (I
think) on a bigger, faster, 64-bit windoze box. I've also discovered more
linux-windoze inconsistencies but that's another story for another day..

-alex 








********************************DISCLAIMER****************************
The information contained in the above e-mail message or messages 
(which includes any attachments) is confidential and may be legally 
privileged.  It is intended only for the use of the person or entity 
to which it is addressed.  If you are not the addressee any form of 
disclosure, copying, modification, distribution or any action taken 
or omitted in reliance on the information is unauthorised.  Opinions 
contained in the message(s) do not necessarily reflect the opinions 
of the Queensland Government and its authorities.  If you received 
this communication in error, please notify the sender immediately 
and delete it from your computer system network.

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/users/attachments/20110609/425fdac4/attachment.html>


More information about the Users mailing list