There is a GNU/R toolbox for the purpose.

http://cran.r-project.org/package=RoughSets

http://cran.r-project.org/web/packages/RoughSets/RoughSets.pdf

Hi Mani A.

Recently, I often use the toolbox you mentioned. But I found that it consumes much time when using fuzzy-rough set. To learn 3000 instances and classify 1000 instances with 14 features, it spends nearly 10 minutes. Other classifier like kNN, SVM, Bayes, Decision tree (also run on R studio) that I tried need just some seconds. Do you have any ideas? Is that problem of rough set or problem of the package?

Obviously a coding issue. File a bug report including details.

Is there any matlab toolbox?

You can use ROSETTA

A Rough Set Toolkit for Analysis of Data

http://www.lcb.uu.se/tools/rosetta/

or maybe you can use ROSE2

http://idss.cs.put.poznan.pl/site/55.html

ROSETTA is very good tool...You can also try the following:

http://cran.r-project.org/web/packages/RoughSets/index.html

Try also, our version of weka at: https://dl.dropboxusercontent.com/u/2043486/weka.jar

which has fuzzy-rough as well as rough methods implemented.

Another to try is KEEL at:

http://sci2s.ugr.es/keel/links.php

Although not as easy to use as weka, it does offer some additional functionality in terms of statistical testing.

ROSETTA is a better tool

Is there a matlab toolbox

toolbox for roughset analysis is ROSETTA, RSES, ROSE, RSTUDIO