|
Arc Add-Ons
Add-ons are files that can be read by
Arc and will add new features to the program. These files end
in ".lsp" or ".fsl", and should be put
in your Extras directory, and the file will be then be
loaded automatically whenever you start Arc.
To download and save the file, hold down the mouse button (the right mouse
button on Windows) and select Save this link as, or Download to disk,
depending on your browser. Your browser my add a suffix ".txt" to the file
name when saving; if so, you need to rename the file deleting the ".txt".
If
you decide you don't want to use this code any more, simply move it to another
directory, or change its name so it does not end in ".lsp" (or ".fsl").
Most add-ons also have documentation, either as an html document or in a pdf
file.
- LaTeX graphs
-
Translate almost any Arc graph into LaTeX, which you can
then convert to PostScript for high-quality printed graphs! This will work on
Linux/Unix, Mac and Windows, but it requires familiarity with LaTeX.
- Windows interface to MS Excel (updated April 2009)
-
For Windows only, you can add an item to the Tools menu in
Microsoft Excel to
read a spreadsheet, or part of a spreadsheet into
Arc. (If you use Macintosh or Windows, the method described in Appendix
Section A.5.3 of Applied Regression Including Computing and Graphics
can be used to save a spreadsheet as an Arc datafile.)
- Interface between Arc and SAS
-
Andrzej Galecki and Yiqun Zhang from the University of Michigan have produced
examples of how data sets can be exchanged between Arc and SAS.
Get their package
SAS2ARC from their
website. They provide documentation and SAS computer code.
- Normal dispersion models
- Get the file dispmod.lsp for
Windows or Linux/Unix or for
Macintosh
and get the documentation in
dispmod.pdf
to figure out what it does.
- Overdisperse binomials
- The document overdispersion.pdf
describes how to use Arc to fit overdisperse binomial data using
Williams' (1982) method. The data file orobanche.lsp for
Windows or Linux/Unix or for
Macintosh
is used as an example in this document.
- Overdisperse Poisson (March 21, 2000, by Luca Scrucca)
- The document poisson-extra-var.pdf
describes how to use Arc to fit overdisperse Poisson data.
Get the necessary code for
Windows or Linux/Unix or for
Macintosh
The data file salmonellaTA98.lsp for
Windows or Linux/Unix or for
Macintosh
is used as an example in this document.
- ARES plots
- ARES plots using Arc.
- Inverse Gaussian generalized linear models
- Inverse Gaussian generalized linear
models using Arc.
- Mulitnomial logit models; recode dialog
-
John Fox at McMaster University has written several additional
extensions, including: (1) a dialog for recoding variables; (2) fitting
multinomial logistic regression models; (3) a spreadsheet for viewing data
(modified from work by Forrest Young). You can get documentation and the
add-ons from
http://socserv.mcmaster.ca/jfox/Misc/Arc/index.html.
- Nonparametric lack-of-fit test for 2D plots (rev. May 15, 2000; rev
Oct. 3, 2000 to allow for point deletions; bug-fix 9/1/2005)
- This add-on implements a nonparametric lack-of-fit test for a simple or
polynomial regression model on a 2D graph, using a bootstrap method suggested
by J. Hart (1997), Nonparametric Smoothing and Lack-of-Fit Test, New
York: Springer.
Get the version of nptest.lsp for
Windows or Linux/Unix or for
Macintosh
and get the documentation in
nptest.pdf
- Simple cross-validation (November 16, 1999, sligtly revised and
simplified, January 31, 2002)
- This add-on allows you to divide your dataset into two parts, a
construction set, and a validation set. Models are fit to the construction
set, and selected fit statistics are displayed for the construction and
validation sets separately.
Get the version of cv.lsp for
Windows or Linux/Unix or for
Macintosh
and get the documentation.
- Improved Help dialog (March 20, 2000, by Luca Scrucca)
- This add-on provides an
improved the help facility for Arc. Get the documentation in
helpdlg.pdf, and the the code in helpdlg.lsp for
Windows or Linux/Unix or for
Macintosh
- Yeo-Johnson transformation family (Feb, 2001)
- Yeo and Johnson (2000, Biometrika, 954-959) have recently suggested a
family of transformations that can be used in place of the Box-Cox family
when a variable is not strictly postitive. Get the documentation in
yjpower.pdf, and the the code in yjpower.lsp for
Windows or Linux/Unix or for
Macintosh
- Delta method for estimating nonlinear functions of parameters
- As of Version 1.06, the delta method is a standard part of Arc, and you do not need an add-on. The documentation is still available as
DeltaMethod.pdf.
- More add ons!
- Luca Scrucca has produced several additional add-ons at the
University of Perugia in Italy.
Contact us!
arc@stat.umn.edu
Last modified: April 29, 2009
|