[R] fitting regression models

Prof Brian Ripley ripley at stats.ox.ac.uk
Thu Oct 22 17:37:26 CEST 1998

> From: root <mop41059 at mail.telepac.pt>

[apparently from Jorge M. A. Magalhaes]

> I have downloaded the package modreg for R but when i doing  "R INSTALL 
>  I get_ errors.

Probably due to the lack of a Fortran compiler: we need more infomation
on those errors.  Please send fuller details to R-devel (and I guess you
have a Linux box and I do not, so I will not be able to help you much).

I should say that modreg is not a public, supported package but a
test-release of code I have written intended for incorporation into R.
I was not aware that its existence was known to R-help: I only
sent information (once) to R-devel asking for comments on installation etc.

> With the packages akima, funfits, acepack,KernSmooth and othes i have the same
> problem. Why??

Probably because you have no functional Fortran compiler: all those and modreg
use Fortran.

> I need some packages for fitting nonlinear models.

Not much in R as yet: you could try minimizing the sum of squares or
maximizing the likelihood directly with nlm().

> I would like install thes fuctions: rreg(); lmsfit(); l1fit(); and loess.
> How i can download and install it?

Those are S-PLUS functions (except lmsfit: you probably mean lmsreg)
and are not part of R.  I know of no implementation of l1fit for R, nor do
think it would be worthwhile to make one.

loess is part of modreg.

rreg and lmsreg are provided as compatibility functions in my experimental
package lqs at http://www.stats.ox.ac.uk/pub/R. As this only uses C
you may be able to use it.  Note that I think the main functions there,
rlm and lqs, are much to be preferred to rreg and lmsreg, both in terms of the
statistics and in terms of the interfaces.

PLEASE: modreg and lqs are experimental, so feedback direct to me or to
R-devel only. In particular, lqs is not finalised yet.

A final comment: if you want so much of the `plus' of S-PLUS, why not
use it?  As I understand it, it is not a design goal of R to clone the
`plus', and there are serious copyright and other issues in trying to
do so.  (loess and rreg are part of S, AFAIK.)

Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list