W dniu 03.03.2011 00:36, Allin Cottrell pisze:
The procedure you're suggesting is not a test (I mean, there's no
null hypothesis), so it probably should not go under the Tests
Fact, there is no hypothesis, but together with Tadeusz Kufel we stated that such feature would be in 'Test' menu, but even better place id Edit/Modify model.
menu. And I'm not sure that creation of a (possibly large) set of
single-observation dummies is very helpful to the GUI user. Might
it not be more helpful to offer something under Edit/Modify model,
along the lines of "Exclude observations with residuals greater
than <select N> * sigma in absolute value"?
The value 2.2 is taken from N(0,1): this is approx two-tailed 3% (exact value is 2.17). But in GUI there could be - instead of critical value - \alpha and critical taken from N(0,1). The idea is not excluding obs (which would be interesting) but creating dummies for i.e. looking for structural breaks or just shocks.
Such dummies are helpful in model selection procedures, proper identifying AR(p) structure and so on.

Regards,
Marcin


Allin

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


-- 
Marcin Błażejowski
http://www.wrzosy.nsb.pl/~marcin/
GG# 203127