R help - Jan 2004 - individual likelihoods

Dear all,
is there a way to extract individual likelihoods from a glm/lrm object?
By individual likelihoods, I mean the likelihoods whose product give the
overall likelihood of the model.
I guess the code in the base package, used to compute the Akaike Information
Criterion may help me.
However, I couldn't figure it out, probably because I'm rather new to
likelihood theory and ML estimation ;-)
Thanks for any help/suggestion/tip,
    Bruno

Dear Bruno,

residuals(mod, type="deviance") gives you the deviance residuals for
the
model, so sum(residuals(mod, type="deviance")^2) is the deviance.

I hope that this helps,
  John

At 02:29 PM 1/16/2004 +0100, Bruno Giordano wrote:
>Dear all,
>is there a way to extract individual likelihoods from a glm/lrm object?
>By individual likelihoods, I mean the likelihoods whose product give the
>overall likelihood of the model.
>I guess the code in the base package, used to compute the Akaike Information
>Criterion may help me.
>However, I couldn't figure it out, probably because I'm rather new
to
>likelihood theory and ML estimation ;-)
>Thanks for any help/suggestion/tip,
>     Bruno
-----------------------------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario, Canada L8S 4M4
email: jfox at mcmaster.ca
phone: 905-525-9140x23604
web: www.socsci.mcmaster.ca/jfox

> 
> Dear all,
> is there a way to extract individual likelihoods from a glm/lrm object?
> By individual likelihoods, I mean the likelihoods whose product give the
> overall likelihood of the model.
> I guess the code in the base package, used to compute the Akaike
Information
> Criterion may help me.
> However, I couldn't figure it out, probably because I'm rather new
to
> likelihood theory and ML estimation ;-)
The aic function just sums the corresponding density ("d") function
with log=T (except for the normal and inverse Gauss, where it is written out
explicitly). Note that mle are not available for the dispersion
parameter of the gamma and inverse Gauss, although this makes vary
little difference in almost all cases. Thus, you just need to feed the
fitted values (and, if appropriate, the dispersion estimate) into the
corresponding density function without summing.
  Cheers, Jim
> Thanks for any help/suggestion/tip,
>     Bruno
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
>