<html>
<head>
</head>
<body>
<br>
<br>
<div class="moz-text-flowed" style="font-family: courier; font-size: 13px; "><br>
We are glad to announce the publication of a new book <br>
<br>
*************************************************************************
<br>
<br>
J.A.K. Suykens, T. Van Gestel, J. De Brabanter, B. De Moor, J. Vandewalle,
<br>
Least Squares Support Vector Machines, <br>
World Scientific Pub. Co., Singapore, 2002 <br>
<a class="moz-txt-link-freetext" href="http://www.esat.kuleuven.ac.be/sista/lssvmlab/book.html">
http://www.esat.kuleuven.ac.be/sista/lssvmlab/book.html</a>
<br>
<br>
*************************************************************************
<br>
<br>
<br>
This book focuses on Least Squares Support Vector Machines (LS-SVMs)
<br>
which are reformulations to standard SVMs. LS-SVMs are closely related
<br>
to regularization networks and Gaussian processes but additionally <br>
emphasize and exploit primal-dual interpretations from optimization theory.
<br>
The authors explain the natural links between LS-SVM classifiers and
kernel <br>
Fisher discriminant analysis. Bayesian inference of LS-SVM models is
<br>
discussed, together with methods for imposing sparseness and employing
<br>
robust statistics. <br>
<br>
The framework is further extended towards unsupervised learning by <br>
considering PCA analysis and its kernel version as a one-class modelling
<br>
problem. This leads to new primal-dual support vector machine formulations
<br>
for kernel PCA and kernel CCA analysis. Furthermore, LS-SVM formulations
<br>
are given for recurrent networks and control. In general, support vector
<br>
machines may pose heavy computational challenges for large data sets.
<br>
For this purpose, a method of fixed size LS-SVM is proposed where the
<br>
estimation is done in the primal space in relation to a Nyström sampling
<br>
with active selection of support vectors. The methods are illustrated
<br>
with several examples. <br>
<br>
<br>
Contents: <br>
. Introduction <br>
. Support vector machines <br>
. Least squares support vector machines, links with Gaussian <br>
processes, regularization networks, and kernel FDA <br>
. Bayesian inference for LS-SVM models <br>
. Weighted versions and robust statistics <br>
. Large scale problems: Nystrom sampling, reduced set methods, <br>
basis formation and Fixed size LS-SVM <br>
. LS-SVM for unsupervised learning: support vector machines <br>
formulations for kernel PCA. Related methods of kernel CCA. <br>
. LS-SVM for recurrent networks and control <br>
. Illustrations and applications <br>
<br>
<br>
Readership: <br>
Graduate students and researchers in neural networks; machine learning;
<br>
data-mining; signal processing; circuit, systems and control theory;
<br>
pattern recognition; and statistics. <br>
<br>
<br>
Info: 308pp., Publication date: Nov. 2002, <br>
ISBN 981-238-151-1 <br>
<br>
Order information: World Scientific <br>
<a class="moz-txt-link-freetext" href="http://www.wspc.com/books/compsci/5089.html">
http://www.wspc.com/books/compsci/5089.html</a>
<br>
<a class="moz-txt-link-freetext" href="http://www.esat.kuleuven.ac.be/sista/lssvmlab/book.html">
http://www.esat.kuleuven.ac.be/sista/lssvmlab/book.html</a>
<br>
<br>
<br>
<br>
******************************************************** <br>
<br>
LS-SVMlab: <br>
Least Squares - Support Vector Machines Matlab/C Toolbox <br>
<a class="moz-txt-link-freetext" href="http://www.esat.kuleuven.ac.be/sista/lssvmlab/">
http://www.esat.kuleuven.ac.be/sista/lssvmlab/</a>
<br>
<br>
******************************************************** <br>
<br>
Toolbox: <br>
. Matlab LS-SVMlab1.4 - Linux and Windows Matlab/C code <br>
. Basic and advanced versions <br>
. Functional and object oriented interface <br>
<br>
<br>
Tutorial User's Guide (100pp.): <br>
. Examples and demos <br>
. Matlab functions with help <br>
<br>
<br>
Solving and handling: <br>
. Classification, Regression <br>
. Tuning, cross-validation, fast loo, <br>
receiver operating characteristic (ROC) curves <br>
. Small and unbalanced data sets <br>
. High dimensional input data <br>
. Bayesian framework with three levels of inference <br>
. Probabilistic interpretations, error bars <br>
. hyperparameter selection, automatic relevance determination (ARD) <br>
input selection, model comparison <br>
. Multi-class encoding/decoding <br>
. Sparseness <br>
. Robustness, robust weighting, robust cross-validation <br>
. Time series prediction <br>
. Fixed size LS-SVM, Nystrom method, <br>
kernel principal component analayis (kPCA), ridge regression <br>
. Unsupervised learning <br>
. Large scale problems <br>
<br>
<br>
Related links, publications, presentations and book: <br>
<a class="moz-txt-link-freetext" href="http://www.esat.kuleuven.ac.be/sista/lssvmlab/">
http://www.esat.kuleuven.ac.be/sista/lssvmlab/</a>
<br>
<br>
<br>
Contact: <a class="moz-txt-link-abbreviated" href="mailto:LS-SVMlab@esat.kuleuven.ac.be">
LS-SVMlab@esat.kuleuven.ac.be</a>
<br>
<br>
<br>
GNU General Public License: <br>
The LS-SVMlab software is made available for research purposes only <br>
under the GNU General Public License. LS-SVMlab software may not be <br>
used for commercial purposes without explicit written permission after <br>
contacting <a class="moz-txt-link-abbreviated" href="mailto:LS-SVMlab@esat.kuleuven.ac.be">
LS-SVMlab@esat.kuleuven.ac.be</a>
. <br>
<br>
[we apologize for receiving multiple copies of this message] <br>
<br>
<br>
<br>
</div>
<br>
<br>
<br>
<br>
<br>
<br>
</body>
</html>