Re: [Gretl-users] transfer function noise models
On Wed, 11 Mar 2009, Allin Cottrell wrote:
On Wed, 11 Mar 2009, bob mccall wrote:
> I just tried Gretl (1.8.0-cvs, for windows) out again and it's
> an amazing program! It's very easy to use. I'm using the Arima
> modeling routine for arima and armax models. I thought that
> Gretl could estimate transfer function noise models as well. But
> Gretl won't allow me to add? lagged dependant variables to the
> list of explanatary variables in the Arima model screen. Am I
> doing something wrong?
Hmm, I'll let Jack Lucchetti or Ignacio Diaz-Emparanza answer
that one.
A similar request was made on the user list about a year ago. You'll find
the thread that ensued at
http://lists.wfu.edu/pipermail/gretl-users/2008-February/thread.html
In short, gretl doesn't offer at the moment a command for estimating
models of the form
y_t = A(L)/B(L) x_t + C(L)/D(L) u_t
natively. Of course, you can write a script that does that via mle or nls,
but I suspect it's not trivial (though feasible).
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti