[R] Converting S-Plus Libraries to R

[(Ted Harding)]

On 06-Jul-04 schimpanski at gmx.de wrote:
> Dear all!
> 
> I'd like to do multiple imputation of missing values with s-plus
> libraries that are provided by Shafer
> (http://www.stat.psu.edu/~jls/misoftwa.html).
> I wonder, whether these libraries are compatible or somehow convertible
> to R (because I don't have S-plus), so that I can use this functions
> using the R Program.
> 
> I would be happy if you could tell me,
> -if it is possible to use S-plus libraries with R
> -if yes, how I can use the S-Plus libraries in R
> 
> Thank you very much,
> 
> Will

Schafer's multiple imputation libraries for S-Plus, on his website,
are CAT, NORM, MIX and PAN.

These have all been converted to R libraries, by various people.
Go to

  http://www.r-project.org

and then, via "CRAN", to any convenient CRAN site, e.g.

  http://cran.at.r-project.org/

and then go to "Packages", which will take you to

  http://cran.at.r-project.org/src/contrib/PACKAGES.html

where you will find a list of all extra libraries. In this list
you will find "cat", "norm", "mix" and "pan", and each of these
will take you to a page for the package in question from which
you can download either the R source of the package or a compiled
binary version for Windows or Mac OX-X, as well as the reference
manual for the package in PDF format which summarises usage and
provides examples of usage with supplied example datasets.

As a general rule, the packages should work exactly as Schafer
intended; any modifications introduced in the R versions are there
either to avoid problems due to differences between R and S-plus,
or to evade bugs or problems which may have been detected in Schafer's
original software.

These packages have been created directly from the S-plus code on
Schafer's website. They have counterparts in the S-plus library
"missing", whose routines are based on Schafer's originals.
For "cat", "norm" and "mix" the corresponding S-plus routines
are in the "Loglin", "Gauss" and "Cgm" [conditional gaussian model]
families of routines in the S-plus "missing" library. I'm not sure
about what corresponds to "pan". Nor am I sure what differences from
Schafer's original public-domain software may have been introduced
when the S-plus library was created.

Hoping this helps,
Ted.


--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at nessie.mcc.ac.uk>
Fax-to-email: +44 (0)870 167 1972
Date: 07-Jul-04                                       Time: 11:56:13
------------------------------ XFMail ------------------------------