[ADMB Users] NLMM Model Selection

[H. Skaug]

The marginal likelihood (= integrated likelhood = likelihood) at the
fit is given in the par file as
"Objective function value =".

Hans


On Tue, Feb 22, 2011 at 4:36 PM, Chris Gast <cmgast at gmail.com> wrote:
> Thank you, Hans.
> Do you know of a way to output the optimum marginal likelihood from ADMB, or
> is this available only internally?
>
> -----------------------------
> Chris Gast
> cmgast at gmail.com
>
>
> On Tue, Feb 22, 2011 at 1:13 AM, H. Skaug <hskaug at gmail.com> wrote:
>>
>> Hi,
>>
>> >
>> > 1) Is it correct to say that the objective function value returned in
>> > the
>> > .par value of a RE model is the approximation to the marginal
>> > likelihood,
>> > post-integration?
>>
>> No
>>
>> > 2a) Is it correct to obtain the conditional likelihood value by
>> > re-evaluating the likelihood function at the MLEs (for fixed parameters)
>> > and
>> > empirical Bayes estimates of REs?
>>
>> Yes, if I read you right.
>>
>> > 2b) If 2a is true, can this value be obtained in a single ADMB
>> > model-fit, or
>> > do I need to first fit the model, then re-run the model using the
>> > optimum
>> > values from the previous run and output an initial likelihood value to
>> > get
>> > the conditional likelihood?
>>
>> Single run. The .par file contains MLEs of fixed effects and empirical
>> bayes
>> estimtes of random effects. The MLEs are approximates, but that does
>> not concern you conceptual disucssion.
>>
>> Hans
>
>