?? osu svm classifier matlab toolbox.htm
字號:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<!-- saved from url=(0045)http://eewww.eng.ohio-state.edu/~maj/osu_svm/ -->
<HTML><HEAD><TITLE>OSU SVM Classifier Matlab Toolbox</TITLE>
<META content="text/html; charset=iso-8859-1" http-equiv=Content-Type>
<META content=en-us http-equiv=Content-Language>
<META content="MSHTML 5.00.3315.2870" name=GENERATOR>
<META content=FrontPage.Editor.Document name=ProgId></HEAD>
<BODY><B><FONT size=+2>OSU SVM Classifier Matlab Toolbox (ver 3.00)</FONT></B>
<P><B><FONT size=+1>By <A href="mailto:junshui@lanl.gov">Junshui Ma</A>, <A
href="mailto:zhaoy@ee.eng.ohio-state.edu">Yi Zhao</A>, and <A
href="http://eewww.eng.ohio-state.edu/~sca">Stanley Ahalt</A></FONT></B>
<P><B>What's new in version 3.00?</B>
<P>(1) The core part of this toolbox is now based on <A
href="http://www.csie.ntu.edu.tw/~cjlin/libsvm/">LIBSVM</A> version 2.33 <BR>(2)
It can be run on Windows, Linux (x86), Solaris, and HP-XP. <BR>(3) Two
data-preprocessing functions are added, <I>Normalize.m</I> and <I>Scale.m</I>.
<BR>
<HR SIZE=1 width="100%">
<BR><B>1. Why we are here?</B>
<P>Although there are quite a few versions of SVM codes in different languages,
including C, Fortran, Java, and Matlab, available on the web, we still found it
is necessary to develop a new Matlab SVM toolbox. Matlab is one of the most
popular research languages in the world. However, currently, the only Matlab SVM
toolbox available is too elementary to deal with any practical problem, which is
due to both its high memory requirement and slow computation speed.
<P>Compared with other available SVM packages, our OSU SVM toolbox has the
following major features:
<P>(1) SVM Classifier implemented as Matlab toolbox enables its users take full
advantage of the convenience of the Matlab platform.
<P>(2) By virtue of Matlab's mex mechanism, this toolbox implements SVM
classifiers in C++ using Chih-Chung Chang and <A
href="http://www.csie.ntu.edu.tw/~cjlin">Chih-Jen Lin</A>'s <A
href="http://www.csie.ntu.edu.tw/~cjlin/libsvm/">LIBSVM algorithm</A> (<A
href="http://www.csie.ntu.edu.tw/~cjlin/libsvm/copyright">COPYRIGHT_libsvm</A>).
Consequently, it is so fast that it is comparable to all the available C-coded
SVM packages in speed. Meanwhile, it inherits the memory-management mechanism of
LIBSVM, and is capable of dealing with practical classification problem with
huge training set.
<P>(3) It is self-contained. That is, it does not request any extra optimization
software packages to support it.
<P>(4) It implements not only the fundamental SVM classification algorithm, but
also two modified formulations, which are named <FONT
face=Symbol>n</FONT>-SVM and one-class SVM classifiers.
<P><B>2. Download</B>
<P><U>Current version:</U> <BR> You can download current version of
this toolbox, along with a demonstration tutorial, from <A
href="http://eewww.eng.ohio-state.edu/~maj/osu_svm/osu_svm3.00.zip">here</A>
(~850KB).
<P><U>Previous versions:</U>
<UL>
<LI>For OSU SVM Toolbox 2.00, you can download this toolbox, along with a
demonstration tutorial, from <A
href="http://eewww.eng.ohio-state.edu/~maj/osu_svm/osu_svm200.zip">here</A>
(~950KB).
<LI>Mr. <A
href="http://www4.csp.ntu.edu.sg:8000/~ejwlu/MyHomePage/LjwHome.htm">Lu
Juwei</A> has successfully transplanted this a previous version (ver 1.10) of
this OSU SVM Toolbox to a UNIX system (Unix System V Release 4.0 or later
versions on SUN ULTRASPARC2 200MHZ). You can download this package from <A
href="http://eewww.eng.ohio-state.edu/~maj/osu_svm/osu_svm110_unix.tar.gz">here</A>
(~397KB).
<LI>Mr. <A href="http://ab.get.to/">Abdul Sayeed</A> has successfully
transplanted this version (ver 2.00) of OSU SVM Toolbox to a UNIX system
(SUN4U, OS Version 5.7, SPARC, SUNW Ultra-5_10). You can download the
corresponding gunzipped mex files from <A
href="http://eewww.eng.ohio-state.edu/~maj/osu_svm/osu_svm200_unix.tar.gz">here</A>
(~17KB). You need to replace those mex files in my package with these mex
files if you want use this Toolbox in the similar system listed above.
</LI></UL>Hope you enjoy it.
<P><B>3. Install and Learn</B>
<P>After downloading the zipped file, you need to
<P>(1) Create a subdirectory in your Matlab's toolbox directory. (For example:
\Matlab\toolbox\osu_svm\)
<P>(2) Unzip the downloaded file into this subdirectory
<P>(3) Add this subdirectory into the search path of your Matlab by "addpath"
command, or by the "file->set path.." menu item, including the subdirectory
of Demo.
<P>(4) Run the command of "demos" from the Matlab command line to learn how to
use this toolbox.
<P><B>Acknowledgments</B>: <SPAN
style="FONT-FAMILY: 'Times New Roman'; FONT-SIZE: 10pt; mso-fareast-font-family: 'Times New Roman'; mso-ansi-language: EN-US; mso-fareast-language: EN-US; mso-bidi-language: AR-SA">This
work is supported by the Mission Research Cooperation under subcontract
SC-1406-98-0001.<SPAN style="mso-spacerun: yes"></SPAN>Additional support has
been provided by the ARL PET program funded by the HPCMO. Also, we appreciate
Dr. Chih-Jen Lin's patient help via email during our developing this
toolbox.</SPAN> <BR> <BR> </P></BODY></HTML>
?? 快捷鍵說明
復制代碼
Ctrl + C
搜索代碼
Ctrl + F
全屏模式
F11
切換主題
Ctrl + Shift + D
顯示快捷鍵
?
增大字號
Ctrl + =
減小字號
Ctrl + -