[Gretl-users] Re: Unknown parameter in time-varying SSM coefficient
Am 03.12.2021 um 12:14 schrieb Paolo Chirico:
Hi,
can I define a time-varying constant in the transition equation as
B.stconst ={mu*B.signt}
where B.signt is a scalar updated by B.timevar_call "update function"?
It is useful to estimate mu, but I'm not sure B.stconst is updated at
each time!
OK, so you must be talking about the state-space modeling apparatus. In
the corresponding chapter (36) of the guide it says under the subheading
"Intercept in the state equation":
(where state equation = transition equation = state transition equation)
"The (r × 1) vector µ can be specified as a bundle member by the name
stconst. Despite its name, this matrix can be stipulated as time
varying, as explained in the next section."
I haven't tried it, but the most natural approach then would seem to me
to include the wanted calculation for updating B.stconst also into the
update function. So your line "B.stconst = {mu*B.signt}" would go into
that function, because that's the intended way to handle time-varying
stuff; presumably it would come after the updating of B.signt (whatever
that is).
HTH,
sven