[Limdep Nlogit List] BoxCox

[Ana I. Sanjuan]

Dear Mikolaj,

Thank you very much for your message. I´m using NLOGIT 4.0.

The code I´m writing is:
boxcox; lhs = prunid
; rhs = qgramo
; rh2 = one, xname0
; model = 4
; lambda = 0 ; theta = 0; MLE $
Is that what you mean?


ana.

Mikołaj Czajkowski escribió:
> Ana I Sanjuan wrote:
>   
>> Dear LIMDEP users,
>>
>> I´m trying to run a Box-Cox regression with possible different 
>> transformations on the dependent and one independent variable, using the 
>> MLE option  (I tried also with grid search) and Model = 4. I initialize 
>> Lambda and Theta, either at 0 or 1, but I get the following error:
>>
>> Error   395: BOXCOX:;LAMBDA=value not found or MODEL=3 & no THETA found.
>>
>> What am I doing wrong?
>>
>> Thank you for any help
>>     
>
> Dear Ana,
>
> There used to be a bug in the Box-Cox model in the old version on 
> NLogit, which was later fixed by prof. Greene. If you sent the code of 
> the command that generates the error message I could make sure that this 
> is it.
>
> Best regards,
>
>   

-- 
------------------------------------------------------------------------
Ana I. Sanjuán
Secretaria de la AEEA

Unidad de Economía Agroalimentaria y de los Recursos Naturales
Centro de Investigación y Tecnología Agroalimentaria de Aragón (C.I.T.A.)
Gobierno de Aragón

Avda Montan~ana, 930
50.059 Zaragoza

tlf.: 976 71 63 49
fax : 976 71 63 35
------------------------------------------------------------------------
General Secretary of the Spanish Association of Agricultural Economics

Research Fellow
Department of Agrofood Economics and Natural Resources
Centre of Agro-food Research and Technology of Aragón (C.I.T.A.)
Government of Aragón

------------------------------------------------------------------------