Topics in Regression
This page
contains a pair of R packages that implements functions for efficient
regression model development.
Strategies
R Packages plgrahics and regr
plgraphics: Graphical tools for efficient and flexible plotting.
A version is available from R-forge.
Use
install.packages("plgraphics", repos="http://R-forge.R-project.org")
regr: The most important function performs fitting of many
diverse regression models and organizes the output in a unified
and useful way.
For some unclear reason, it is not available at R-forge for the time
being. It can be loaded from below.
See also the
package's website
The ideas behind these packages are described in the document
"The R-Function regr and Package regr0 for an Augmented
Regression Analysis"
In case that installing the packages should not work,
here is the linux version
of plgraphics and
of regr,
which you can download and install or install directly by writing
install.packages("http://stat.ethz.ch/~stahel/regression/regr_1.0-5.tar.gz")
and analogously for plgraphics.
For windows, plgraphics is available both on R-forge and
here.
The regr package is
here
R Package lassogrp
A preliminary version of a package for the lasso model selection
procedure is also available from R-forge.
Use
install.packages("lassogrp", repos="http://r-forge.r-project.org",
lib=...)
see also the
package's website
In case that installing the package should not work, here is
the source code, which you can download and
source into your workspace by calling
source("http://stat.ethz.ch/~stahel/regression/lassogrp.R").
Talks
Regression Model Development and Yet Another Regression Function.
Talk given at the useR 2008 conference, Dortmund, Aug 13, 2008
Abstract /
Slides