系统参数辨识的信息准则及算法 英文2025|PDF|Epub|mobi|kindle电子书版本百度云盘下载

系统参数辨识的信息准则及算法 英文PDF格式电子书版下载

注意：本站所有压缩包均有解压码： 点击下载压缩包解压工具

1 Introduction1

1.1 Elements of System Identification1

1.2 Traditional Identification Criteria3

1.3 Information Theoretic Criteria4

1.3.1 MEE Criteria6

1.3.2 Minimum Information Divergence Criteria7

1.3.3 Mutual Information-Based Criteria7

1.4 Organization of This Book8

Appendix A:Unifying Framework of ITL9

2 Information Measures13

2.1 Entropy13

2.2 Mutual Information19

2.3 Information Divergence21

2.4 Fisher Information23

2.5 Information Rate24

Appendix B：α-Stable Distribution26

Appendix C:Proof of(2.17)26

Appendix D:Proof of Cramer-Rao Inequality27

3 Information Theoretic Parameter Estimation29

3.1 Traditional Methods for Parameter Estimation29

3.1.1 Classical Estimation29

3.1.2 Bayes Estimation31

3.2 Information Theoretic Approaches to Classical Estimation34

3.2.1 Entropy Matching Method34

3.2.2 Maximum Entropy Method35

3.2.3 Minimum Divergence Estimation37

3.3 Information Theoretic Approaches to Bayes Estimation40

3.3.1 Minimum Error Entropy Estimation40

3.3.2 MC Estimation51

3.4 Information Criteria for Model Selection56

Appendix E:EM Algorithm57

Appendix F:Minimum MSE Estimation58

Appendix G:Derivation of AIC Criterion58

4 System Identification Under Minimum Error Entropy Criteria61

4.1 Brief Sketch of System Parameter Identification61

4.1.1 Model Structure62

4.1.2 Criterion Function65

4.1.3 Identification Algorithm65

4.2 MEE Identification Criterion72

4.2.1 Common Approaches to Entropy Estimation73

4.2.2 Empirical Error Entropies Based on KDE76

4.3 Identification Algorithms Under MEE Criterion82

4.3.1 Nonparametric Information Gradient Algorithms82

4.3.2 Parametric IG Algorithms86

4.3.3 Fixed-Point Minimum Error Entropy Algorithm91

4.3.4 Kernel Minimum Error Entropy Algorithm93

4.3.5 Simulation Examples95

4.4 Convergence Analysis104

4.4.1 Convergence Analysis Based on Approximate Linearization104

4.4.2 Energy Conservation Relation106

4.4.3 Mean Square Convergence Analysis Based on Energy Conservation Relation111

4.5 Optimization of φ-Entropy Criterion122

4.6 Survival Information Potential Criterion129

4.6.1 Definition of SIP129

4.6.2 Properties of the SIP131

4.6.3 Empirical SIP136

4.6.4 Application to System Identification139

4.7 △-Entropy Criterion143

4.7.1 Definition of △-Entropy145

4.7.2 Some Properties of the △-Entropy148

4.7.3 Estimation of △-Entropy152

4.7.4 Application to System Identification157

4.8 System Identification with MCC161

Appendix H:Vector Gradient and Matrix Gradient164

5 System Identification Under Information Divergence Criteria167

5.1 Parameter Identifiability Under KLID Criterion167

5.1.1 Definitions and Assumptions168

5.1.2 Relations with Fisher Information169

5.1.3 Gaussian Process Case173

5.1.4 Markov Process Case176

5.1.5 Asymptotic KLID-Identifiability180

5.2 Minimum Information Divergence Identification with Reference PDF186

5.2.1 Some Properties188

5.2.2 Identification Algorithm196

5.2.3 Simulation Examples198

5.2.4 Adaptive Infinite Impulsive Response Filter with Euclidean Distance Criterion201

6 System Identification Based on Mutual Information Criteria205

6.1 System Identification Under the MinMI Criterion205

6.1.1 Properties of MinMI Criterion207

6.1.2 Relationship with Independent Component Analysis211

6.1 3 ICA-Based Stochastic Gradient Identification Algorithm212

6.1.4 Numerical Simulation Example214

6.2 System Identification Under the MaxMI Criterion216

6.2.1 Properties of the MaxMI Criterion217

6.2.2 Stochastic Mutual Information Gradient Identification Algorithm222

6.2.3 Double-Criterion Identification Method227

Appendix I:MinMI Rate Criterion238

References239