Part I: Preliminaries 1

Chapter 1. Introduction and Book Overview 3

Organization of the Second Volume 3

Acknowledgements 8

Chapter 2. Methods Used in this Book 9

i. Basic Principles of Classification 9

ii. Main Ideas of the SVM Algorithm for Binary Classification 13

iii. Mathematical Formulation of SVMs 16

iv. Theoretical Underpinnings of SVMs 17

v. Model Selection and Accuracy Estimation 18

vi. Accuracy Metrics 23

vii. Statistical Comparison of Classification Results 24

viii. SVMs for Variable/Feature Selection 27

ix. Other Variable/Feature Selection Methods Used in this Book 30

Part II: Case Studies and Comparative Evaluation in High-Throughput

Genomic Data 33

Chapter 3. Application and Comparison of SVMs and Other Methods

for Multicategory Microarray-Based Cancer Classification 35

Methods 36

Results 43

Conclusions 49

Chapter 4. Comparison of SVMs and Random Forests for Microarray-Based Cancer Classification 51

Methods 52

Results 57

Conclusions 62

Chapter 5. Comparison of SVMs and Kernel Ridge Regression for Microarray-Based Cancer Classification (Contributed by Zhiguo Li) 65

Methods 65

Results 69

Conclusions 72

Chapter 6. Application and Comparison of SVMs and Other Methods for Multicategory Classification in Microbiomics (Contributed by Mikael Henaff, Kranti Konganti, Varun Narendra, Alexander V. Alekseyenko) 73

Methods 74

Results 81

Conclusions 86

Chapter 7. Application to Assessment of Plasma Proteome Stability 87

Methods 87

Results 92

Conclusions 93

Part III: Case Studies and Comparative Evaluation in Text Data 95

Chapter 8. Application and Comparison of SVMs and Other Methods for Retrieving High-Quality Content-Specific Articles (Contributed by Yindalon Aphinyanaphongs) 97

Methods 98

Results 103

Conclusions 105

Chapter 9. Application and Comparison of SVMs and Other Methods for Identifying Unproven Cancertreatments on the Web (Contributed by Yindalon Aphinyanaphongs) 107

Methods 108

Results 111

Conclusions 112

Chapter 10. Application to Predicting Future Article Citations (Contributed by Lawrence Fu) 113

Methods 113

Results 116

Conclusions 121

Chapter 11. Application to Classifying Instrumentality of Article Citations (Contributed by Lawrence Fu) 123

Methods 124

Results 127

Conclusions 130

Chapter 12. Application and Comparison of SVMs and Other Methods for Identifying Drug Drug Interactions-Related Literature (Contributed by Stephany Duda) 131

Methods 131

Results 135

Conclusions 138

Part IV: Case Studies with Clinical Data 139

Chapter 13. Application to Predicting Clinical Laboratory Values 141

Methods 141

Results 147

Conclusions 150

Chapter 14. Application to Modeling Clinical Judgment and Guideline Compliance in the Diagnosis of Melanoma (Contributed by Andrea Sboner) 151

Methods 151

Results 155

Conclusions 160

Part V: Other Comparative Evaluation Studies of Broad Applicability 161

Chapter 15. Using SVMs for Causal Variable Selection 163

Methods 164

Results 167

Conclusions 176

Chapter 16. Application and Comparison of SVM-RFE and GLL Methods for Variable Selection for Classification 177

Methods 178

Results 182

Conclusions 187

Conclusions and Lessons Learned 189

Bibliography 191

Additional Reading for Presented Case Studies and Benchmarks 191

List of References 193

Index 199

中科院文献情报中心