Suchen und Finden

Titel

Autor/Verlag

Inhaltsverzeichnis

Nur eBooks für mein Endgerät anzeigen:

 

Newsletter

Machine Learning - Modeling Data Locally and Globally

Machine Learning - Modeling Data Locally and Globally

von: Kaizhu Huang, Haiqin Yang, Irwin King, Michael Lyu

Springer-Verlag, 2008

ISBN: 9783540794523, 179 Seiten

Format: PDF, OL

Mac OSX,Windows PC Apple iPad, Android Tablet PC's Online-Lesen für: Linux,Mac OSX,Windows PC

Preis: 139,05 EUR

Mehr zum Inhalt

Machine Learning - Modeling Data Locally and Globally


 

Preface

6

Contents

7

1 Introduction

11

1.1 Learning and Global Modeling

11

1.2 Learning and Local Modeling

13

1.3 Hybrid Learning

15

1.4 Major Contributions

15

1.5 Scope

18

References

19

2 Global Learning vs. Local Learning

23

2.1 Problem De.nition

25

2.2 Global Learning

26

2.4 Hybrid Learning

33

2.5 Maxi-Min Margin Machine

34

References

35

3 A General Global Learning Model: MEMPM

39

3.1 Marshall and Olkin Theory

40

3.2 Minimum Error Minimax Probability Decision Hyperplane

41

3.3 Robust Version

55

3.4 Kernelization

56

3.5 Experiments

60

3.6 How Tight Is the Bound?

66

3.7 On the Concavity of MEMPM

70

3.8 Limitations and Future Work

75

3.9 Summary

76

References

77

4 Learning Locally and Globally: Maxi-Min Margin Machine

79

4.1 Maxi-Min Margin Machine

81

4.2 Bound on the Error Rate

92

4.3 Reduction

94

4.4 Kernelization

95

4.5 Experiments

98

4.6 Discussions and Future Work

103

4.7 Summary

103

References

104

5 Extension I: BMPM for Imbalanced Learning

107

5.1 Introduction to Imbalanced Learning

108

5.2 Biased Minimax Probability Machine

108

5.3 Learning from Imbalanced Data by Using BMPM

110

5.4 Experimental Results

112

5.5 When the Cost for Each Class Is Known

124

5.6 Summary

125

References

125

6 Extension II: A Regression Model from M4

129

6.1 A Local Support Vector Regression Model

131

6.2 Connection with Support Vector Regression

132

6.3 Link with Maxi-Min Margin Machine

134

6.4 Optimization Method

134

6.5 Kernelization

135

6.6 Additional Interpretation on wTSiw

137

6.7 Experiments

138

6.8 Summary

141

References

141

7 Extension III: Variational Margin Settings within Local Data in Support Vector Regression

143

7.1 Support Vector Regression

144

7.2 Problem in Margin Settings

146

7.3 General -insensitive Loss Function

146

7.4 Non-.xed Margin Cases

149

7.5 Experiments

151

7.6 Discussions

165

References

168

8 Conclusion and Future Work

171

8.1 Review of the Journey

171

8.2 Future Work

173

References

174

Index

177