Advances in Large Margin Classifiers

Front Cover
Alexander J. Smola
MIT Press, 2000 - 412 pages

The book provides an overview of recent developments in large margin classifiers, examines connections with other methods (e.g., Bayesian inference), and identifies strengths and weaknesses of the method, as well as directions for future research.

The concept of large margins is a unifying principle for the analysis of many different approaches to the classification of data from examples, including boosting, mathematical programming, neural networks, and support vector machines. The fact that it is the margin, or confidence level, of a classification--that is, a scale parameter--rather than a raw training error that matters has become a key tool for dealing with classifiers. This book shows how this idea applies to both the theoretical analysis and the design of algorithms. The book provides an overview of recent developments in large margin classifiers, examines connections with other methods (e.g., Bayesian inference), and identifies strengths and weaknesses of the method, as well as directions for future research. Among the contributors are Manfred Opper, Vladimir Vapnik, and Grace Wahba.

From inside the book

Contents

vi
22
Roadmap
31
Dynamic Alignment Kernels
39
Natural Regularization from Generative Models
51
Probabilities for SV Machines
61
Maximal Margin Perceptron
75
Large Margin Rank Boundaries for Ordinal Regression
115
Generalized Support Vector Machines
135
Towards a Strategy for Boosting Regressors
247
Bounds on Error Expectation for SVM
261
Adaptive Margin Support Vector Machines
281
GACV for Support Vector Machines
297
Mean Field and LeaveOneOut
311
Computing the Bayes Kernel Classifier
329
Margin Distribution and Soft Margin
349
Support Vectors and Statistical Mechanics
359

Linear Discriminant and Support Vector Classifiers
147
Regularization Networks and Support Vector Machines
171
Functional Gradient Techniques for Combining Hypotheses
221
Entropy Numbers for Convex Combinations and MLPs
369
Copyright

Other editions - View all

Common terms and phrases

Popular passages

Page 406 - In B. Scholkopf, CJC Burges, and AJ Smola, editors, Advances in Kernel Methods — Support Vector Learning, pages 69-88, Cambridge, MA, 1999b.
Page 392 - I. Daubechies, Ten Lectures on Wavelets, SIAM, Philadelphia, PA, 1992. Notes from the 1990 CBMS-NSF Conference on Wavelets and Applications at Lowell, MA.
Page 390 - Mar. 1991. [2] M. Bertero, T. Poggio, and V. Torre, "Ill-posed problems in early vision,

About the author (2000)

Alexander J. Smola is Senior Principal Researcher and Machine Learning Program Leader at National ICT Australia/Australian National University, Canberra.

Bibliographic information