Neo/NST Support Vector Machine Package

This package is a collection of several (currently only two) algorithms that utilize so called kernel functions. You can check-out this package from the CVS repository with cvs -co folder_svm.

Some general links about Support Vector Machines (SVM):

Kernel-Machines.org Comprehensive collection of links about kernel methods.

Publications Searchable list of publications on kernel methods (provided by: www.kernel-machines.org).

SVM Tutorial I C.J.C. Burges: A Tutorial on Support Vector Machines for Pattern Recognition.
Knowledge Discovery and Data Mining, 2(2), 1998.

SVM Tutorial II A. J. Smola and B. Schölkopf: A Tutorial on Support Vector Regression.
NeuroCOLT Technical Report NC-TR-98-030, Royal Holloway College,
University of London, UK, 1998.

SVM Book N. Cristianini and J. Shawe-Taylor: An introduction to support vector machines : and other kernel-based learning methods. Cambridge University Press , 2000. (UniBib: HK632 C933)

Further information about the provided Neo/NST units:

unit SMO

Description: Support Vector Machine for binary pattern classification. The unit uses J. Platt's Sequential Minimal Optimization algorithm for adaption.The following kernel types are provided: Radial Basis Function (RBF), polynomial, tanh and linear. Additionally, a generic kernel function can be loaded from a shared object (use this template for coding a generic kernel).

Manpage: SMO

Literature: J. Platt: Fast training of support vector machines using sequential minimal optimization.
In B. Schölkopf, C. J. C. Burges, and A. J. Smola, editors, Advances in Kernel Methods --- Support Vector Learning,
pages 185-208, Cambridge, MA, 1999. MIT Press. [ PDF , PS ]

Files: nst_svm.C, smo.H, smo.C

unit Novelty

Description: This unit estimates the support of the training data. Depending on the parameter \nu and the kernel width \sigma of the RBF-kernel, the unit estimates the hull of the data. Outliers can be handled by chosing 0 < \nu < 1 which determines the number of examples outside of the hull.

Manpages: Novelty

Literature: B. Schölkopf, J. Platt, J. Shawe-Taylor, A. J. Smola, and R. C. Williamson:
Estimating the support of a high-dimensional distribution.
Technical Report 99-87, Microsoft Research, 1999. To appear in Neural Computation, 2001. [ PS ]

Files: nst_novelty.c, CNovelty.H, CNovelty.C,

Note: This unit exploits dual-processor machines.

Send further questions to ttwellma@TechFak.Uni-Bielefeld.de .

Kernel-Machines.org	Comprehensive collection of links about kernel methods.
Publications	Searchable list of publications on kernel methods (provided by: www.kernel-machines.org).
SVM Tutorial I	C.J.C. Burges: A Tutorial on Support Vector Machines for Pattern Recognition. Knowledge Discovery and Data Mining, 2(2), 1998.
SVM Tutorial II	A. J. Smola and B. Schölkopf: A Tutorial on Support Vector Regression. NeuroCOLT Technical Report NC-TR-98-030, Royal Holloway College, University of London, UK, 1998.
SVM Book	N. Cristianini and J. Shawe-Taylor: An introduction to support vector machines : and other kernel-based learning methods. Cambridge University Press , 2000. (UniBib: HK632 C933)

Description:	Support Vector Machine for binary pattern classification. The unit uses J. Platt's Sequential Minimal Optimization algorithm for adaption.The following kernel types are provided: Radial Basis Function (RBF), polynomial, tanh and linear. Additionally, a generic kernel function can be loaded from a shared object (use this template for coding a generic kernel).
Manpage:	SMO
Literature:	J. Platt: Fast training of support vector machines using sequential minimal optimization. In B. Schölkopf, C. J. C. Burges, and A. J. Smola, editors, Advances in Kernel Methods --- Support Vector Learning, pages 185-208, Cambridge, MA, 1999. MIT Press. [ PDF , PS ]
Files:	nst_svm.C, smo.H, smo.C

Description:	This unit estimates the support of the training data. Depending on the parameter \nu and the kernel width \sigma of the RBF-kernel, the unit estimates the hull of the data. Outliers can be handled by chosing 0 < \nu < 1 which determines the number of examples outside of the hull.
Manpages:	Novelty
Literature:	B. Schölkopf, J. Platt, J. Shawe-Taylor, A. J. Smola, and R. C. Williamson: Estimating the support of a high-dimensional distribution. Technical Report 99-87, Microsoft Research, 1999. To appear in Neural Computation, 2001. [ PS ]
Files:	nst_novelty.c, CNovelty.H, CNovelty.C,
Note:	This unit exploits dual-processor machines.