Difference between revisions of "LIBSVM -- A Library for Support Vector Machines"

From Cohen Courses
Jump to navigationJump to search
 
(3 intermediate revisions by the same user not shown)
Line 1: Line 1:
−
The ACE 2005 [[Category::Tool|tool]] addresses five primary tasks – the recognition of entities, values, temporal expressions, [[Relation Extraction|relations]], and events.  
+
LIBSVM is an integrated software for support vector classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and distribution estimation (one-class SVM). It supports multi-class classification. It is actively patched and has interfaces in many different language including Java, Matlab, C# etc. [http://www.csie.ntu.edu.tw/~cjlin/libsvm/]
  
−
The dataset is available at the Linguistic Data Consortium. The data is taken from a variety of sources and is available for the tasks in the following languages: Arabic, Chinese and English.
+
Since version 2.8, it implements an SMO-type algorithm proposed in this paper:
 +
R.-E. Fan, P.-H. Chen, and C.-J. Lin. Working set selection using second order information for training SVM. Journal of Machine Learning Research 6, 1889-1918, 2005.  
  
−
Four versions of each document are provided:
+
Main features of LIBSVM include
−
* Source text files (.sgm): All source files, including the Chinese files, are encoded in UTF-8.
+
* Different SVM formulations
−
* APF files (.apf.xml): The ACE Program Format.
+
* Efficient multi-class classification
−
* AG files (.ag.xml): The LDC Annotation Graph Format.
+
* Cross validation for model selection
−
* TABLE files (.tab): Files that store mapping tables between the IDs used in each ag.xml file and their corresponding
+
* Probability estimates
−
apf.xml file.
+
* Weighted SVM for unbalanced data
 +
* Both C++ and Java sources
 +
* GUI demonstrating SVM classification and regression
 +
* Python, R (also Splus), MATLAB, Perl, Ruby, Weka, Common LISP, CLISP, Haskell and LabVIEW interfaces. C# .NET code is available. It's also included in some data mining environments: RapidMiner and PCP.
 +
* Automatic model selection which can generate contour of cross valiation accuracy.
  
−
The detailed statistics for the training portion of this corpus are as follows:
+
Its website maintains a list of packages for download and includes a practical guide for beginner to start with [http://www.csie.ntu.edu.tw/~cjlin/papers/guide/guide.pdf].
−
 
 
−
[[File:ACE05-1.png]]
 
−
 
 
−
[http://www.itl.nist.gov/iad/mig//tests/ace/2005/ External Link]
 
−
 
 
−
{{#ask: [[UsesDataset::ACE 2005 dataset]]
 
−
| ?AddressesProblem
 
−
| ?UsesDataset
 
−
}}
 

Latest revision as of 23:38, 30 September 2010

LIBSVM is an integrated software for support vector classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and distribution estimation (one-class SVM). It supports multi-class classification. It is actively patched and has interfaces in many different language including Java, Matlab, C# etc. [1]

Since version 2.8, it implements an SMO-type algorithm proposed in this paper: R.-E. Fan, P.-H. Chen, and C.-J. Lin. Working set selection using second order information for training SVM. Journal of Machine Learning Research 6, 1889-1918, 2005.

Main features of LIBSVM include

  • Different SVM formulations
  • Efficient multi-class classification
  • Cross validation for model selection
  • Probability estimates
  • Weighted SVM for unbalanced data
  • Both C++ and Java sources
  • GUI demonstrating SVM classification and regression
  • Python, R (also Splus), MATLAB, Perl, Ruby, Weka, Common LISP, CLISP, Haskell and LabVIEW interfaces. C# .NET code is available. It's also included in some data mining environments: RapidMiner and PCP.
  • Automatic model selection which can generate contour of cross valiation accuracy.

Its website maintains a list of packages for download and includes a practical guide for beginner to start with [2].

Retrieved from "http://curtis.ml.cmu.edu/w/courses/index.php?title=LIBSVM_--_A_Library_for_Support_Vector_Machines&oldid=1956"