[R-sig-Geo] Bayesian model comparison

Roger Bivand Roger.Bivand at nhh.no
Thu Jul 16 23:52:16 CEST 2015


On Thu, 16 Jul 2015, Antonio Bubbico wrote:

> Hello Roger,
>
> thanks for that. Is there any available code that I can take as an example?

INLABMA is on CRAN; maybe see the documentation for the slm latent model 
om the R-INLA website, and possibly a paper in the JSS spatial statistics 
special number.

Roger

>
> Regards.
>
> 2015-07-16 11:41 GMT+02:00 Roger Bivand <Roger.Bivand at nhh.no>:
>
>> On Thu, 16 Jul 2015, Antonio Bubbico wrote:
>>
>>  Hello,
>>>
>>> I would like to make a bayesian model comparison among SLX, SDM and SDEM
>>> and produce model probabilities for all three. I would like to make it
>>> rather than using likelihood
>>> ratio tests that rely on two-way comparisons (Le Sage, 2014).
>>>
>>> Can you suggest to me some package to make the model comparison among the
>>> three models?
>>>
>>
>> Experimentally, try INLABMA and INLA with the slm latent model. The slm
>> latent model itself with BIC or similar may be enough. These estimation
>> methods are still under development.
>>
>> Roger
>>
>>
>>> Thanks.
>>>
>>>
>>>
>> --
>> Roger Bivand
>> Department of Economics, Norwegian School of Economics,
>> Helleveien 30, N-5045 Bergen, Norway.
>> voice: +47 55 95 93 55; fax +47 55 95 91 00
>> e-mail: Roger.Bivand at nhh.no
>>
>>
>
>
>

-- 
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 91 00
e-mail: Roger.Bivand at nhh.no



More information about the R-sig-Geo mailing list