Skip to content

Quantile GAM?

3 messages · Jonathan Greenberg, Gavin Simpson, Mark Difford

#
R-ers:

    I was wondering if anyone had suggestions on how to implement a GAM 
in a quantile fashion?  I'm trying to derive a model of a "hull" of 
points which are likely to require higher-order polynomial fitting (e.g. 
splines)-- would quantreg be sufficient, if the response and predictors 
are all continuous?  Thanks!

--j
#
On Fri, 2009-05-29 at 16:51 -0700, Jonathan Greenberg wrote:
If you're just after the AM bit of GAM, then yes, I think quantreg
contains functions to do as you wish. See ?rqss and ?qss for models that
allow smoothing splines in quantile regression.

Alternatively, I think package VGAM may also contain something of use.

HTH

G
#
Hi Jonathan,
Take a look at the gamlss package.

Regards, Mark.
Jonathan Greenberg-2 wrote: