Xuegong Zhang,
Xin Lu
and Wing H. Wong
Institute of Bioinformatic
/ Dept. of Automation , Tsinghua
University,
Beijing 100084, China
Department of Statistics,
Harvard University, Cambridge, MA
02138, USA
Department of Biostatistics,
Harvard School of Public Health,
Boston, MA 02115, USA
New version is available for downloading now at the end of this page.
Methodology Description:
R-SVM is a SVM-based method for doing supervised pattern
recognition(classification) with microarray gene expression data.
The method uses SVM for both classification and for selecting a subset
of relevant genes according to their relative contribution in the classification.
This process is done recursively so that a series of gene subsets and classification
models can be obtained in a recursive manner, at different levels of gene
selection. The performance of the classification can be evaluated
either on an independent test data set or by cross validation on the same
data set. R-SVM also includes an option for permutation experiments
to assess the significance of the performance.
Detailed description of the methodology and procedures
can be found in:
ZHANG, X.G., LU, X., (Joint First Author) XU, X.Q., LEUNG, H.E., WONG, W.H. and LIU, J.S. (2006) RSVM: A SVM based Strategy for Recursive Feature Selection and Sample Classification with Proteomics Mass-Spectrometry Data. BMC Bioinformatics, 7:197
Software Package:
This package includes the binary files for R-SVM
with the cross validation and permutation test procedures described in
the technical report. The package was compiled to run under Red Hat
Linux release 6.2. Some other versions will be available later at
this same website. Before that, for users of other systems, please
contact Xuegong Zhang to
check the availability.
See the files README.txt, USAGE.txt and EXAMPLE.txt
coming with the package for general instructions, usage instructions and
examples.
I'd like to remind you that if you want to use R-SVM,
you'd better feel comfortable with using Linux/UNIX systems, although
you don't need to be an expert at all. The reason that we do not use
the more popular platforms is that some of the computations might take
hours, days or even longer time, depending on your specific computer and
data, it is more convenient (and safer!) to run such big jobs on UNIX/Linux
environment. Sorry if this causes extra effort on your side.
Whoever downloads this package is regarded as a
"user" of R-SVM. By downloading this package, the user agrees that
the downloaded package will be used for academic use only, and agrees that
any work based on or directly related to the download software and/or its
documentations will cite the original work by Xuegong Zhang and Wing H.
Wong (see the readme.txt file for details). We should also be informed
for any such publications. For potential commercial users, please
contact Xuegong Zhang or
Wing H. Wong before using
it on real data.
Anyone who downloads this package is not supposed
to circulate it without the agreement of the authors. Please refer
to this website for anyone who is interested in obtaining the package.
The package is provided as is. We are working
hard to make it correct and in good shape, but we hereby disclaim all warranties
with regard to this program, including merchantability and fitness.
Users are obligated to report any bugs in the package, either technical
or methodological.
This package utilizes the SVMTorch package (Ronan
Collobert and Samy Bengio,
SVMTorch: support vector machines for large-scale regression problems,
Journal
of Machine Learning Research, vol.1, pp.143-160, 2001 http://www.idiap.ch/learning/SVMTorch.html).
Users of R-SVM are required to take the responsibility to follow the original
license agreement of SVMTorch. (The license file of SVMTorch also
comes with the R-SVM package).
Click Here to Download R code of R-SVM, written by R language, and use R package e1071
Note on 12/31/05: This is the new version written in R by Xin Lu of HSPH.. Please report any bugs or inconveniences.