[Gretl-users] Re: Question about the Kalman Filter - MLE combination
Am 22.05.2024 um 21:36 schrieb Alecos Papadopoulos:
I understand that simple non-negative constraints can be imposed in
other indirect ways, but I have to impose bounds on the estimated
parameters like "0<alpha<1".
This comment is not directly about
your syntax question, but interval
constraints like this can also be imposed indirectly. A typical way of
doing that is to map the parameter to the unit interval by using a CDF
like the standard normal or so. Of course, the bounds of the interval
can be adjusted as well. Sorry if this is beside the point, but I
thought it wouldn't hurt to mention it.
So can I somehow combine the naked mle "check" syntax with the
KF-success syntax into one, say
Again, not directly about your question, but it is allowed to call
user-defined functions in an mle block. So you can define your own
function that takes the relevant series and parameters, and spits out
either a scalar result or perhaps also another series which may also
contain missings. This approach is much more flexible than using a
single conditional statement, so I guess it could help solve your problem.
cheers
sven