Re: [Gretl-devel] Proposal for function package cleanup
El 16/06/15 a las 19:56, Riccardo (Jack) Lucchetti escribió:
> HEGY_test by Lucchetti is older, and GHegy tried to improve it
> (obtaining not only 95% critival values but also p-values) and
> generalize it for any seasonal periodicity. Then HEGY_test may be
> considered deprecated if Jack agrees.
No problem, let's just get rid of mine.
The only thing in partial disagreement that I feel like saying,
however, is: competing implementation of a given technique may be nice
to have (the standard bio-diversity metaphor applies). But certainly,
only up to a point. My implementation hasn't been maintained for a
long time and it should go.
I agree in that we should admit competing implementations of a given
technique.
In this particular case if HEGY_test were including p-values it could be
expected to be more precise than GHegy for the specific case S=4,
because the first one is fitting a surface-reponse regresion for this
specific value of S=4, and the second is fitting a surface-response
regresion to differente values of S. But I must say that I did some
checking about the 10%, 5% and 1% critical values finding that the
differences between the two procedures were in general not superior to
+-0.01
--
Ignacio Díaz-Emparanza
Departamento de Economía Aplicada III (Econometría y Estadística)
Universidad del País Vasco - Euskalherriko Unibertsitatea, UPV/EHU
Tfno: (+34) 94 601 3732
http://www.ehu.eus/ea3