📄 mysvm.htm
字号:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<!-- saved from url=(0048)http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/ -->
<!-- header mySVM --><HTML><HEAD><TITLE>mySVM</TITLE>
<META content="text/html; charset=gb2312" http-equiv=Content-Type>
<META content="MSHTML 5.00.2920.0" name=GENERATOR></HEAD>
<BODY bgColor=#ffffff><A name=top></A><!--- Header table --->
<TABLE bgColor=#eeeeee border=0 cellPadding=0 cellSpacing=0 width="100%">
<COLGROUP>
<COL width="20%">
<COL width="70%">
<COL width="10%"></COLGROUP>
<TBODY>
<TR>
<TD rowSpan=2 width="20%"><A
href="http://www-ai.cs.uni-dortmund.de/logo.html"><IMG border=0
src="mySVM.files/eier_graybg.gif"></A></TD>
<TD bgColor=#99cdff width="70%"><A href="http://www.uni-dortmund.de/"
target=_top><IMG alt="University of Dortmund" border=0
src="mySVM.files/balken_le.gif"></A></TD>
<TD align=right bgColor=#eeeeee width="10%"><A
href="http://www.uni-dortmund.de/" target=_top><IMG alt=UniDo-Logo
border=0 src="mySVM.files/balken_ro.gif"></A></TD></TR>
<TR>
<TD><A href="http://www.cs.uni-dortmund.de/" target=_top><IMG
alt="Computer Science" border=0 src="mySVM.files/cs.gif"></A> <A
href="http://www-ai.cs.uni-dortmund.de/" target=_top><IMG
alt="Artificial Intelligence" border=0 src="mySVM.files/ai.gif"></A></TD>
<TD align=right vAlign=top><A href="http://www.uni-dortmund.de/"
target=_top><IMG alt=UniDo-Logo border=0
src="mySVM.files/balken_ru.gif"></A></TD></TR>
<TR>
<TD bgColor=#eeeeee colSpan=3 height=5> <!--- spacer trick --></TD></TR><!--- XX button row -->
<TR>
<TD bgColor=#99cdff colSpan=2 height=25 noWrap vAlign=bottom><A
href="http://www-ai.cs.uni-dortmund.de/index.eng.html"
onmousedown="imgClick('img_ls8news'); return true"
onmouseout="imgNormal('img_ls8news'); return true"
onmouseover="imgOver('img_ls8news'); return true"><IMG alt="LS8 News"
border=0 height=25 name=img_ls8news src="mySVM.files/ls8news.gif"
width=77></A> <A
href="http://www-ai.cs.uni-dortmund.de/FORSCHUNG/index.eng.html"
onmousedown="imgClick('img_forschung'); return true"
onmouseout="imgNormal('img_forschung'); return true"
onmouseover="imgOver('img_forschung'); return true"><IMG alt=Research
border=0 height=25 name=img_forschung src="mySVM.files/forschung.eng.gif"
width=82></A><A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/index.eng.html"
onmousedown="imgClick('img_software'); return true"
onmouseout="imgNormal('img_software'); return true"
onmouseover="imgOver('img_software'); return true"><IMG alt=Software
border=0 height=25 name=img_software src="mySVM.files/software.gif"
width=83></A><A
href="http://www-ai.cs.uni-dortmund.de/PARTNER/index.eng.html"
onmousedown="imgClick('img_partner'); return true"
onmouseout="imgNormal('img_partner'); return true"
onmouseover="imgOver('img_partner'); return true"><IMG alt=Partner
border=0 height=25 name=img_partner src="mySVM.files/partner.gif"
width=73></A> <A
href="http://www-ai.cs.uni-dortmund.de/LEHRE/lehre.html"
onmousedown="imgClick('img_lehre'); return true"
onmouseout="imgNormal('img_lehre'); return true"
onmouseover="imgOver('img_lehre'); return true"><IMG alt=Teaching border=0
height=25 name=img_lehre src="mySVM.files/lehre.eng.gif" width=79></A>
<A
href="http://www-ai.cs.uni-dortmund.de/PERSONAL/personal.eng.html"
onmousedown="imgClick('img_personal'); return true"
onmouseout="imgNormal('img_personal'); return true"
onmouseover="imgOver('img_personal'); return true"><IMG alt=Staff border=0
height=25 name=img_personal src="mySVM.files/personal.eng.gif"
width=54></A><A
href="http://www-ai.cs.uni-dortmund.de/UNIVERSELL/index.eng.html"
onmousedown="imgClick('img_allgemein'); return true"
onmouseout="imgNormal('img_allgemein'); return true"
onmouseover="imgOver('img_allgemein'); return true"><IMG alt=General
border=0 height=25 name=img_allgemein src="mySVM.files/allgemein.eng.gif"
width=74></A><A href="http://www-ai.cs.uni-dortmund.de/INTERN/intern.html"
onmousedown="imgClick('img_intern'); return true"
onmouseout="imgNormal('img_intern'); return true"
onmouseover="imgOver('img_intern'); return true"><IMG alt=Internal
border=0 height=25 name=img_intern src="mySVM.files/intern.eng.gif"
width=85></A></TD>
<TD align=right bgColor=#99cdff height=25 noWrap vAlign=bottom><A
href="http://www-ai.cs.uni-dortmund.de/Harvest/brokers/www-ai/query.eng.html"
onmousedown="imgClick('img_search'); return true"
onmouseout="imgNormal('img_search'); return true"
onmouseover="imgOver('img_search'); return true"><IMG alt=Search border=0
height=25 name=img_search src="mySVM.files/search.gif" width=30></A><A
href="mailto:webadmin@ls8.cs.uni-dortmund.de"
onmousedown="imgClick('img_mail'); return true"
onmouseout="imgNormal('img_mail'); return true"
onmouseover="imgOver('img_mail'); return true"><IMG
alt="Send email to webadmin@ls8.cs.uni-dortmund.de" border=0 height=25
name=img_mail src="mySVM.files/mail.gif" width=38></A><IMG alt="no german"
border=0 height=25 src="mySVM.files/no_deutsch.gif"
width=32></TD></TR></TBODY></TABLE>
<SCRIPT language=JavaScript src="mySVM.files/buttons.eng.js"
type=text/javascript></SCRIPT>
<!--- Body table -->
<TABLE width="100%">
<TBODY>
<TR>
<TD colSpan=3><FONT size=1> </FONT></TD></TR>
<TR>
<TD> </TD>
<TD>
<H1>mySVM</H1><!-- /header -->
<CENTER>
<H1>mySVM - a support vector machine</H1>by <A
href="http://www-ai.cs.uni-dortmund.de/PERSONAL/rueping.html">Stefan
Rüping</A>, <A
href="mailto:rueping@ls8.cs.uni-dortmund.de">rueping@ls8.cs.uni-dortmund.de</A>
</CENTER>
<H2>News </H2>
<UL>
<LI>Download the latest release of <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/mySVM-latest.tar.gz">mySVM</A>
(Version 2.1.1, November 7th, 2001)
<LI>Download the <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/mySVM-latest-bin.zip">binary
version for Windows</A>
<LI>See a <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/changes.eng.html">list
of changes</A> </LI></UL>
<H2>About mySVM </H2>mySVM is an implementation of the Support Vector
Machine introduced by V. Vapnik (see <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#Vapnik/98a">[Vapnik/98a]</A>).
It is based on the optimization algorithm of <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/SVM_LIGHT/svm_light.eng.html">SVM<I><SUP>light</SUP></I></A>
as described in <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#Joachims/99a">[Joachims/99a]</A>.
mySVM can be used for pattern recognition, regression and distribution
estimation.
<H2>License </H2>This software is free only for non-commercial use. It
must not be modified and distributed without prior permission of the
author. The author is not responsible for implications from the use of
this software.
<P>If you are using mySVM for research purposes, please cite the software
manual available from this cite in your publications (Stefan Rüping
(2000): <EM>mySVM-Manual</EM>, University of Dortmund, Lehrstuhl
Informatik 8, http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/).
<H2>Installation </H2>
<H3>Installation under Unix</H3>
<UL>
<LI>Download <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/mySVM-latest.tar.gz">mySVM</A>.
<LI>Create a new directory, change into it and unpack the files into
this directory
<LI>On typical UN*X systems simply type <TT>make</TT> to compile mySVM.
On other systems you have to call your C++ compiler manually. </LI></UL>If
everything went right you should have a new subdirectory named
<TT>bin</TT> and to files <TT>mysvm</TT> and <TT>predict</TT> in a
subdirectory thereof. On some systems you might get an error message about
<TT>sys/times.h</TT>. If you do, open the file <TT>globals.h</TT> and
uncomment the line <TT>#undef use_time</TT>.
<H3>Installation under Windows</H3>If you get the <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/mySVM-latest.tar.gz">source
code</A> version, you have to compile mySVM youself. First edit the file
<EM>globals.h</EM> and uncomment the line <TT>#define windows 1</TT>.
Compile the file <EM>learn.cpp</EM> to get the learning program and
<EM>predict.cpp</EM> for the model application program. mySVM was tested
under Visual C++ 6.0. You can also get the <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/mySVM-latest-bin.zip">binary
version</A>. <A name=usage>
<H2>Using mySVM </H2></A>For a complete reference of mySVM have a look
into the mySVM manual (<A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/mysvm-manual.ps">Postscript</A>,
<A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/mysvm-manual.pdf">PDF</A>).
Here is a short users guide:
<UL>
<LI><TT>mysvm</TT> is used for training a SVM on a given example set and
testing the results
<LI><TT>predict</TT> is used for predicting the functional value of new
examples based on an already trained SVM. </LI></UL>The input of mySVM
consists of
<UL>
<LI>a <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#paramdef">parameter
definition</A>
<LI>a <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#kerneldef">kernel
definition</A>
<LI>one or more <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#exampledef">example
sets</A> </LI></UL>Input lines starting with "#" are treated as
commentary. The input can be given in one or more files. If no filenames
or the filename "-" are given, the input is read from stdin.
<TT>mysvm</TT> trains a SVM on the first given example set. The following
example sets are used for testing (if their classification is given) or
the functional value of the examples is being computed (if no
classification is given). <A name=paramdef>
<H3>Parameter definition</H3></A>The parameter definition lets the user
choose the type of loss function, the optimizer parameters and the
training algorithm to use. The parameter definition starts with the line
<TT>@parameters</TT>.
<H4>Global parameters:</H4>
<TABLE border=1>
<TBODY>
<TR>
<TD>pattern</TD>
<TD>use SVM for pattern recognition</TD></TR>
<TR>
<TD>regression</TD>
<TD>use regression SVM <EM>(default)</EM></TD></TR>
<TR>
<TD>nu <EM>float</EM></TD>
<TD>use nu-SVM with the given value of nu instead of normal SVM (see
<A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#Schoelkopf/etal/2000a">[Schoelkopf/etal/2000a]</A>
for details on nu-SVMs).
<TR>
<TD>distribution</TD>
<TD>estimate the support of the distribution of the training
examples (see <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#schoelkopf/etal/99a">[Schoelkopf/etal/99a]</A>).
Nu must be set!
<TR>
<TD>verbosity [1..5]</TD>
<TD>ranges from 1 (no messages) over 3 (default) to 5 (flood, for
debugging only) </TD></TR>
<TR>
<TD>scale</TD>
<TD>scale the training examples to mean 0 and variance 1
(default)</TD></TR>
<TR>
<TD>no_scale</TD>
<TD>do not scale the training examples (may be numerically less
stable!)</TD></TR>
<TR>
<TD>format</TD>
<TD>set the default example file format. See the description <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#exampledef">here</A>.</TD></TR>
<TR>
<TD>delimiter</TD>
<TD>set the default example file format. See the description <A
href="http://www-ai.cs.uni-dortmund.de/SOFTWARE/MYSVM/#exampledef">here</A>.</TD></TR></TBODY></TABLE>
<H4>Loss function:</H4>
<TABLE border=1>
<TBODY>
<TR>
<TD>C <EM>float</EM></TD>
<TD>the SVM complexity constant (Note: C will be scaled by 1 /
number of training examples).</TD></TR>
<TR>
<TD>L+ <EM>float</EM></TD>
<TD>penalize positive deviation (prediction too high) by this
factor</TD></TR>
<TR>
<TD>L- <EM>float</EM></TD>
<TD>penalize negative deviation (prediction too low) by this
factor</TD></TR>
<TR>
<TD>epsilon <EM>float</EM></TD>
<TD>insensitivity constant. No loss if prediction lies this close to
true value</TD></TR>
<TR>
<TD>epsilon+ <EM>float</EM></TD>
<TD>epsilon for positive deviation only</TD></TR>
<TR>
<TD>epsilon- <EM>float</EM></TD>
<TD>epsilon for negative deviation only</TD></TR>
<TR>
<TD>quadraticLoss+</TD>
<TD>use quadratic loss for positive deviation</TD></TR>
<TR>
<TD>quadraticLoss-</TD>
<TD>use quadratic loss for negative deviation</TD></TR>
<TR>
<TD>quadraticLoss</TD>
<TD>use quadratic loss for both positive and negative
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -