[Gretl-users] Re: gretl questions- quantile regression
On Mon, 15 Dec 2008, Deborah wrote:
I have a few questions for you. I'm trying to analyze a data set
using quantile regression. Your software is incredibly easy to
use and offers a lot of great options for this kind of analysis.
But I'd like to do a non-linear quantile regression? Is this
possible?
Sorry, there's no provision for nonlinear quantile regression in
gretl at present.
As a second question, is there a way, using gretl, to test for
heteroskedasticity in order to decide which options are best for
the regression? Thanks for the help (and the great program!),
Do you mean, in the context of a quantile regression? If so,
there's no built-in functionality of that sort. For OLS you have
White's test and the Breusch-Pagan test, and for IV regression you
have the HET1 test (Pesaran and Taylor). These are documented
under the help for "lmtest".
Allin Cottrell