Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st:about the maximum dimension of mixlogit estimation


From   Nina YIN <[email protected]>
To   [email protected]
Subject   Re: st:about the maximum dimension of mixlogit estimation
Date   Tue, 19 Jun 2012 15:33:27 +0200

Hi Maarten,

Do you mean I should set id* variables as fixed coefficients? That makes sense.
But I guess I might need to estimate more than 10 additional random
effects, do you think it's impossible?
Thank you very much for your comments!

Best,
Nina

On Tue, Jun 19, 2012 at 3:22 PM, Maarten Buis <[email protected]> wrote:
> -mixlogit- fits a random effects model, and you just do not want to
> estimate such a model with more than 20 random effects. It is hard
> enough to estimate a model with one or two random effects. So that
> default makes perfect sense to me.
>
> -- Maarten
>
> On Tue, Jun 19, 2012 at 3:17 PM, Nina YIN <[email protected]> wrote:
>> Hi Nick,
>>
>> Thank you for your reminding.
>>
>> mixlogit is a user written command programmed by Prof. Arne Risa Hole
>> which fits mixed logit models by using maximum simulated likelihood.
>>
>> The command I typed in is
>> -mixlogit choice if avail==1,group(ncs) id(np) rand(id* X1 X2 )
>>
>> where id* are the product-specific variables, X1 X2 are the
>> interaction terms of alternative characteristics and individual
>> attributes.
>> ncs is choice situation identifier, np is individual identifier.
>>
>> Then the result returns error with message that maximum dimension is 20.
>>
>>
>>
>>
>> On Tue, Jun 19, 2012 at 1:40 PM, Nick Cox <[email protected]> wrote:
>>> This lacks explanation of where -mixlogit- comes from and the exact code that you typed.
>>>
>>> See Statalist FAQ if you are mystified on why these are expected.
>>>
>>> Nick
>>> [email protected]
>>>
>>> Nina YIN
>>>
>>> I am trying to estimate a mixed logit model using command -mixlogit.
>>> Because of the endogeneity problem, I include  product-specific fixed
>>> effects(around 20 products) in rand() option to control the
>>> unobservable products characteristics, and also I include some
>>> interaction terms of alternative-characteristics with individual
>>> attributes in rand() option(not much, at the beginning I tried 2 or 3
>>> interaction terms). Then the error comes out with" maximum dimension
>>> is 20".
>>>
>>> Does that mean the maximum number of coefficients can be no larger
>>> than 20? What determines the maximum dimension in this model?
>>> In one of my reference, the author use -clogit fixed coefficient logit
>>> model, the number of his coefficients is around 200. Why in my model,
>>> I can't replicate it using -mixlogit command?
>>>
>>>
>>> *
>>> *   For searches and help try:
>>> *   http://www.stata.com/help.cgi?search
>>> *   http://www.stata.com/support/statalist/faq
>>> *   http://www.ats.ucla.edu/stat/stata/
>>
>>
>>
>> --
>> Best Regards,
>> Nina YIN
>>
>>
>> Toulouse School of Economics
>> Manufacture des Tabacs
>> 21 Allee de Brienne
>> Toulouse, 31000, France
>> Tel: 0033-(0)5 6123 8348
>>
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/statalist/faq
>> *   http://www.ats.ucla.edu/stat/stata/
>
>
>
> --
> --------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
>
> http://www.maartenbuis.nl
> --------------------------
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/



-- 
Best Regards,
Nina YIN


Toulouse School of Economics
Manufacture des Tabacs
21 Allee de Brienne
Toulouse, 31000, France
Tel: 0033-(0)5 6123 8348

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index