On Sat, 25 Jun 2011, savos schmagges wrote:

Dear gretl-community,

is it possible to impose restrictions inside a mle-command?
e.g. i want to impose that a parameter d should be estimated higher or equal to a parameter
b.
so i am searching for something like that:

mle logl=...
...
d>=b
params d b
end mle

Is there any possibility to account for that??

Not directly. However, to achieve what you want you may reparametrise your model in such a way that the problem becomes unconstrained. Example: suppose you have 2 parameters, b and d, and that you want to ensure that d>b. You may define a scalar g such that d = b + exp(g) and maximise your loglikelihood in the b-g space, as in

<hansl>
mle logl = ...
   ...
   d = b + exp(g)
   params b g
end mle
</hansl>

Of course, you should choose the initial value for g in a sensible way.

Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche

r.lucchetti@univpm.it
http://www.econ.univpm.it/lucchetti

_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users