Dear Jack,
I'm still stuck in bivariate probit models and suspect something strange going out there.
When I estimate a (highly correlated) bivariate model (providing the --robust and --cluster options) throught the GUI and try to save the $yhat matrix through the console, I got the same (not so) error message:
> The statistic you requested is not available
>>> genr series predict_external = $yhat[,1]
Instead, when I estimate the model through the console (or script) everything works fine
???
Best
Artur
 
 
 
From: Riccardo (Jack) Lucchetti
Date: 2014-04-28 07:49
To: Gretl list
Subject: Re: [Gretl-users] bivariate probit in a loop
On Mon, 28 Apr 2014, artur.bala.tn wrote:
 
> Dear all,
 
> I'm currently estimating a bootstraped bivariate probit through a
> progressive loop and retrieve each time the $yhat matrix. At some point,
> the execution is interrupted with the "warning":
>
>
> The statistic you requested is not available
>>> genr series predict_external = $yhat[,1]
>
>
> Is there a perfect prediction symptom behind this message isn't it?
 
It could be, yes.
 
> Can one, in such a loop, skip cases where the MLE estimation is not
> technically possible?
 
Use "catch": for example,
 
<hansl>
nulldata 50
 
small = 0.25
x1 = normal()
x2 = normal()
y1 = x1 + normal()*small > 0.5
y2 = x2 > 0
 
catch biprobit y1 y2 const x1 ; const x2
err = $error
if !err
     z = $yhat
     printf "%10.3f", meanc(z)
else
     printf "Error!\n"
endif
</hansl>
 
 
-------------------------------------------------------
   Riccardo (Jack) Lucchetti
   Dipartimento di Scienze Economiche e Sociali (DiSES)
 
   Università Politecnica delle Marche
   (formerly known as Università di Ancona)
 
   r.lucchetti@univpm.it
   http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
 
 
_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users