[Gretl-users] maximum like. conditions

Davor Horvatic davorh at phy.hr
Sun Nov 29 09:24:29 EST 2009


OK, I figured out the correct code, so please confirm that I interpret 
correctly.
Restriction criteria has to be written as

scalar check = (alpha>0) && (beta>0)

or in my case as

scalar check = (alpha>0) && (beta>0) && (gamma1>0)

if mle does not find parameters that are consistent with that, I'll get 
output
like this

-----------------------------------
Tolerance = 1,81899e-012
Function evaluations: 417
Evaluations of gradient: 58

Model 1: ML, using observations 1-12059
ll = check ? -0.5*(log(h) + (e^2)/h) : NA
Standard errors based on Outer Products matrix

                estimate     std. error   t-ratio   p-value
   --------------------------------------------------------
   mu         0,000277590        NA         NA        NA
   omega      0,0989841          NA         NA        NA
   alpha      0,198343           NA         NA        NA
   beta       0,155879           NA         NA        NA
   gamma1     9,09556e-010       NA         NA        NA

Log-likelihood       5542,439   Akaike criterion    -11074,88
Schwarz criterion   -11037,89   Hannan-Quinn        -11062,47
---------------------------------

if this is correct please correct section 17.2

All the best,
Davor



On 29.11.2009. 14:02, Davor Horvatic wrote:
> OK, I can confirm this, that code runs now but constraint is not
> applied, for example
>
> scalar checka = (alpha>0)
> scalar checkb = (beta>0)
> scalar checkc = (gamma1>0)
> scalar check = checka+checkb+checkc
>
> applied to my model gives
>
>                  estimate     std. error   t-ratio     p-value
>     -----------------------------------------------------------
>     mu          0,000115904   0,00314962    0,03680   0,9706
>     omega       0,0991130     0,00457944   21,64      7,07e-104 ***
>     alpha       0,199138      0,0100301    19,85      1,02e-087 ***
>     beta        0,157351      0,0323213     4,868     1,13e-06  ***
>     gamma1     -1,78166       2,06923      -0,8610    0,3892
>
> I did not supply whole code and data because I want to stick to
> avaliable data in the gretle.
>
> If I use only for example scalar check = (alpha>0)
> I get:
>
> Data types not conformable for operation
>
>
>
> On 29.11.2009. 13:26, Davor Horvatic wrote:
>    
>> I'll try immediately, thanks. But still question remains why this
>> does not work in original form. I took code from
>> the end of section 17.2.
>>
>> Thanks once more,
>> Davor
>>
>> On 29.11.2009. 13:06, Berend Hasselman wrote:
>>
>>      
>>> On 29-11-2009, at 12:27, Davor Horvatic wrote:
>>>
>>>
>>>
>>>        
>>>> ? mle ll = check ? (-0.5*(log(h) + (e^2)/h))
>>>> ? series e = y - mu
>>>> ? series h = var(y)
>>>> ? series h = omega + alpha*(e(-1))^2 + beta*h(-1)
>>>> ? scalar check = (alpha>0)&    (beta>0)
>>>> ? params mu omega alpha beta
>>>> ? end mle
>>>> Syntax error in command line
>>>>
>>>> Error executing script: halting
>>>>
>>>>
>>>>          
>>>>> end mle
>>>>>
>>>>>
>>>>>            
>>> A.
>>> ---------------------------------------------
>>> I replaced the&    in the scalar check = .. line  with&&
>>> and got this
>>>
>>> ? mle ll = check ? (-0.5*(log(h) + (e^2)/h)) : NA
>>> ? series e = y - mu
>>> ? series h = var(y)
>>> ? series h = omega + alpha*(e(-1))^2 + beta*h(-1)
>>> ? scalar check = (alpha>0)&&    (beta>0)
>>> ? params mu omega alpha beta
>>> ? end mle
>>> Data types not conformable for operation
>>>
>>> Error executing script: halting
>>>
>>>
>>>        
>>>> end mle
>>>>
>>>>
>>>>          
>>> B.
>>> ----------------------------------------------
>>> I replaced the line scalar check ... by the following three lines
>>>
>>> scalar checka = (alpha>0)
>>> scalar checkb = (beta>0)
>>> scalar check = checka+checkb
>>>
>>> and now the script runs without error to completion.
>>>
>>> Berend
>>> _______________________________________________
>>> Gretl-users mailing list
>>> Gretl-users at lists.wfu.edu
>>> http://lists.wfu.edu/mailman/listinfo/gretl-users
>>>
>>>
>>>        
>> _______________________________________________
>> Gretl-users mailing list
>> Gretl-users at lists.wfu.edu
>> http://lists.wfu.edu/mailman/listinfo/gretl-users
>>
>>      
> _______________________________________________
> Gretl-users mailing list
> Gretl-users at lists.wfu.edu
> http://lists.wfu.edu/mailman/listinfo/gretl-users
>    


More information about the Gretl-users mailing list