Laura: Part of the issue may depend on what you mean by goodness-of-ft.
If you are looking for some global measure like a pseudo R or AIC to
select among models, you ought to be able to make those calculations off
the objective function that was minimized as you recognized. If
qr.fit.sfn() does not return the objective function like rq.fit.br(), the
simplex routine, you still ought to be able to do the calculations by
performing the asymmetric weighting of the residuals from the model (see
Roger Koenker's 2005 book). Now, if by goodness-of-fit you mean how the
model fits in local regions of the predictor space, then you might want to
check out Stef van Buuren's work on worm plots to diagnose fit in quantile
regression. Don't remember where he published this but his email is
Stef.vanBuuren@tno.nl
Brian
Brian S. Cade, PhD
U. S. Geological Survey
Fort Collins Science Center
2150 Centre Ave., Bldg. C
Fort Collins, CO 80526-8818
email: brian_cade@usgs.gov
tel: 970 226-9326
From:
"laura m." <mayoralaura@gmail.com>
To:
r-help@r-project.org
Date:
05/22/2009 03:29 AM
Subject:
[R] Goodness of fit in quantile regression
Sent by:
r-help-bounces@r-project.org
Dear R users,
I've used the function qr.fit.sfn to estimate a quantile regression on a
panel data set. Now I would like to compute an statistic to measure the
goodness of fit of this model.
Does someone know how could I do that? I could compute a pseudo R2 but in
order to do that I would need the value of the objetive function at the
optimum and I don't see how to get this from the function qr.fit.sfn.
If someone has any good idea about how to solve this problem I would be
most
grateful!
Best
Laura
--
View this message in context:
http://www.nabble.com/Goodness-of-fit-in-quantile-regression-tp23666962p23666962.html
Sent from the R help mailing list archive at Nabble.com.
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
[[alternative HTML version deleted]]