Knowledge Discovery with Support Vector Machines

by Lutz H. Hamel

• ISBN13: 9780470503041
• ISBN10: 0470503041
• Format: eBook
• Copyright: 2009-10-01
• Publisher: Wiley-Interscience

Summary

An easy-to-follow introduction to support vector machinesThis book provides an in-depth, easy-to-follow introduction to support vector machines drawing only from minimal, carefully motivated technical and mathematical background material. It begins with a cohesive discussion of machine learning and goes on to cover:Knowledge discovery environmentsDescribing data mathematicallyLinear decision surfaces and functionsPerceptron learningMaximum margin classifiersSupport vector machinesElements of statistical learning theoryMulti-class classificationRegression with support vector machinesNovelty detectionComplemented with hands-on exercises, algorithm descriptions, and data sets, Knowledge Discovery with Support Vector Machines is an invaluable textbook for advanced undergraduate and graduate courses. It is also an excellent tutorial on support vector machines for professionals who are pursuing research in machine learning and related areas.

Table of Contents

Preface

Acknowledgments

What is Knowledge Discovery?

Machine Learning

The Structure of the Universe X

Inductive Learning

Model Representations

Exercises

Bibliographic Notes

Knowledge Discovery Environments

Computational Aspects of Knowledge Discovery

Data Access

Visualization

Data Manipulation

Model Building and Evaluation

Model Deployment

Other Toolsets

Exercises

Bibliographic Notes

Describing Data Mathematically

From Data Sets to Vector Spaces

Vectors

Vector Spaces

The Dot Product as a Similarity Score

Lines, Planes, and Hyperplanes

Exercises

Bibliographic Notes

Linear Decision Surfaces and Functions

From Data Sets to Decision Functions

Linear Decision Surfaces through the Origin

Decision Surfaces with an Offset Term

A Simple Learning Algorithm

Discussion

Exercises

Bibliographic Notes

Perceptron Learning

Perceptron Architecture and Training

Duality

Discussion

Exercises

Bibliographic Notes

Maximum Margin Classifiers

Optimization Problems

Maximum Margins

Optimizing the Margin

Quadratic Programming

Discussion

Exercises

Bibliographic Notes

Support Vector Machines

The Lagrangian Dual

Dual MaximumMargin Optimization

The Dual Decision Function

Linear Support Vector Machines

Non-Linear Support Vector Machines

The Kernel Trick

Feature Search

A Closer Look at Kernels

Soft-Margin Classifiers

The Dual Setting for Soft-Margin Classifiers

Tool Support

WEKA

R

Discussion

Exercises

Bibliographic Notes

Implementation

Gradient Ascent

The Kernel-Adatron Algorithm

Quadratic Programming

Chunking

Sequential Minimal Optimization

Discussion

Exercises

Bibliographic Notes

Evaluating What has been Learned

Performance Metrics

The Confusion Matrix

Model Evaluation

The Hold-Out Method

The Leave-One-Out Method

N-Fold Cross-Validation

Error Confidence Intervals

Model Comparisons

Model Evaluation in Practice

WEKA

R

Exercises

Bibliographic Notes

Elements of Statistical Learning Theory

The VC-Dimension and Model Complexity

A Theoretical Setting for Machine Learning

Empirical Risk Minimization

VC-Confidence

Structural Risk Minim

Table of Contents provided by Publisher. All Rights Reserved.