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.
Author(s): Alexander J. Smola, Peter Bartlett, Bernhard Schölkopf, Dale Schuurmans (Editors)
Series: Advances in Neural Information Processing Systems
Edition: 1st
Publisher: The MIT Press
Year: 2000
Language: English
Commentary: Hyperlinks. Searchable with all (tested) readers. No cover.
Pages: 422
Preface......Page 9
Introduction to Large Margin Classifiers......Page 11
Roadmap......Page 41
Support Vector Machines......Page 47
Dynamic Alignment Kernels......Page 49
Natural Regularization from Generative Models......Page 61
Probabilities for SV Machines......Page 71
Maximal Margin Perceptron......Page 85
Large Margin Rank Boundaries for Ordinal Regression......Page 125
Kernel Machines......Page 143
Generalized Support Vector Machines......Page 145
Linear Discriminant and Support Vector Classifiers......Page 157
Regularization Networks and Support Vector Machines......Page 181
Boosting......Page 215
Robust Ensemble Learning......Page 217
Functional Gradient Techniques for Combining Hypotheses......Page 231
Towards a Strategy for Boosting Regressors......Page 257
Leave-One-Out Methods......Page 269
Bounds on Error Expectation for SVM......Page 271
Adaptive Margin Support Vector Machines......Page 291
GACV for Support Vector Machines......Page 307
Gaussian Processes and SVM: Mean Field and Leave-One-Out......Page 321
Beyond the Margin......Page 337
Computing the Bayes Kernel Classifier......Page 339
Margin Distribution and Soft Margin......Page 359
Support Vectors and Statistical Mechanics......Page 369
Entropy Numbers for Convex Combinations and MLPs......Page 379
References......Page 397
Index......Page 399
