[ADMB Users] Objective function value in robust regression

[Rubén Roa]

> -----Mensaje original-----
> De: users-bounces at admb-project.org 
> [mailto:users-bounces at admb-project.org] En nombre de dave fournier
> Enviado el: miércoles, 12 de mayo de 2010 10:28
> Para: users at admb-project.org
> Asunto: Re: [ADMB Users] Objective function value in robust regression
> 
> 
>   Yes it is the log-likelihood value. What else would it be?

Since ADMB does minimisation, I asked: "does the robust regression procedure reports the _raw negative_ log-likelihood in the par file? The obvious alternative is that it reports the log-likelihood. I am getting positive objective function values (and therefore negative AICs) but we are dealing here with probability densities, so the log-likelihood could be positive or negative. I checked the ADMB manual about the robust regression procedure, it looks cool, with the mixture of distributions, but I didn't find any reference to what it is reported as the result of the value for the objective function. I guess it is the raw negative log-likelihood that was minimised but I would like to have confirmation.

Rubén

____________________________________________________________________________________ 

Dr. Rubén Roa-Ureta
AZTI - Tecnalia / Marine Research Unit
Txatxarramendi Ugartea z/g
48395 Sukarrieta (Bizkaia)
SPAIN