Computational Network Theory Theoretical Foundations and

Applications Theoretical Foundations and Applications, 1st

Edition

Visit the link below to download the full version of this book:

https://medipdf.com/product/computational-network-theory-theoretical-foundations
-and-applications-theoretical-foundations-and-applications-1st-edition/

Click Download Now

Related Titles

Siddique, N., Adeli, H. Junker, B. H., Schreiber, F.

Computational Intelligence Analysis of Biological

Synergies of Fuzzy Logic, Neural Networks
Networks and Evolutionary Computing
2008
2013 ISBN: 978-0-470-04144-4
Hardcover
ISBN: 978-1-118-33784-4
Engelbrecht, A. P.

Dehmer, M., Basak, S. C. Computational Intelligence

An Introduction
Statistical and Machine
Learning Approaches for 2007
Network Analysis ISBN: 978-0-470-03561-0

2012 Dunne, R. A.
ISBN: 978-0-470-19515-4
A Statistical Approach to
Okabe, A., Sugihara, K. Neural Networks for Pattern
Spatial Analysis Along Recognition
Networks 2007
Statistical and Computational Methods ISBN: 978-0-471-74108-4

2012
ISBN: 978-0-470-77081-8
Quantitative and Network Biology
Series Editors M. Dehmer and F. Emmert-Streib
Volume 5

Computational Network Theory

Theoretical Foundations and Applications

Edited by
Matthias Dehmer, Frank Emmert-Streib, and Stefan Pickl
The Editors All books published by Wiley-VCH are
carefully produced. Nevertheless, authors,
Prof. Dr. Matthias Dehmer editors, and publisher do not warrant the
UMIT –The Health and Lifesciences information contained in these books,
University including this book, to be free of errors.
Eduard Wallnoefer Zentrum 1 Readers are advised to keep in mind that
A-6060 Hall in Tirol statements, data, illustrations, procedural
Austria details or other items may inadvertently
be inaccurate.
and
Library of Congress Card No.: applied for

Universität der Bundeswehr München

Institute for Theoretical Computer British Library Cataloguing-in-Publication
Science, Mathematics and Data
Operations Research A catalogue record for this book is
Department of Computer Science available from the British Library.
Werner-Heisenberg-Weg 39
85577 Neubiberg Bibliographic information published by the
Germany Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek
Prof. Dr. Frank Emmert-Streib lists this publication in the Deutsche
Tampere University of Technology Nationalbibliografie; detailed
Department of Signal Processing bibliographic data are available on the
Korkeakoulunkatu 1 Internet at <http://dnb.d-nb.de>.
FI-33720 Tampere
Finland © 2015 Wiley-VCH Verlag GmbH & Co.
KGaA, Boschstr. 12, 69469 Weinheim,
Germany
Prof. Dr. Stefan Pickl
Universität der Bundeswehr München
Institute for Theoretical Computer All rights reserved (including those of
Science, Mathematics and translation into other languages). No part
Operations Research of this book may be reproduced in any
Department of Computer Science form – by photoprinting, microfilm,
Werner-Heisenberg-Weg 39 or any other means – nor transmitted
85577 Neubiberg or translated into a machine language
Germany without written permission from the
publishers. Registered names, trademarks,
etc. used in this book, even when not
specifically marked as such, are not to be
considered unprotected by law.

Print ISBN: 978-3-527-33724-8

ePDF ISBN: 978-3-527-69153-1
ePub ISBN: 978-3-527-69154-8
Mobi ISBN: 978-3-527-69152-4
oBook ISBN: 978-3-527-69151-7

Typesetting Laserwords Private Limited,

Chennai, India
Printing and Binding Markono Print
Media Pte Ltd, Singapore

Printed on acid-free paper

 VII

The all-important glue that holds our wonderful complex societies together is
cooperation.
Jerry Sabloff, President Santa Fe Institute

Dedicated to the 30th birthday of the Santa Fe Institute and the 85th
birthday of Murray Gell-Mann, one of its founders.

What is especially striking and remarkable is that in fundamental physics a

beautiful or elegant theory is more likely to be right than a theory that is
inelegant.
Murray Gell-Mann.
 IX

Contents

Color Plates XV
Preface XXXI
List of Contributors XXXIII

1 Model Selection for Neural Network Models: A Statistical

Perspective 1
Michele La Rocca and Cira Perna
1.1 Introduction 1
1.2 Feedforward Neural Network Models 2
1.3 Model Selection 4
1.3.1 Feature Selection by Relevance Measures 6
1.3.2 Some Numerical Examples 10
1.3.3 Application to Real Data 12
1.4 The Selection of the Hidden Layer Size 14
1.4.1 A Reality Check Approach 15
1.4.2 Numerical Examples by Using the Reality Check 16
1.4.3 Testing Superior Predictive Ability for Neural Network
Modeling 19
1.4.4 Some Numerical Results Using Test of Superior Predictive
Ability 21
1.4.5 An Application to Real Data 23
1.5 Concluding Remarks 26
References 26

2 Measuring Structural Correlations in Graphs 29

Ziyu Guan and Xifeng Yan
2.1 Introduction 29
2.1.1 Solutions for Measuring Structural Correlations 31
2.2 Related Work 32
2.3 Self Structural Correlation 34
2.3.1 Problem Formulation 34
2.3.2 The Measure 34
2.3.2.1 Random Walk and Hitting Time 35
X Contents

2.3.2.2 Decayed Hitting Time 36

2.3.3 Computing Decayed Hitting Time 37
2.3.3.1 Iterative Approximation 37
2.3.3.2 ̃ , B) 39
A Sampling Algorithm for h(v i
2.3.3.3 Complexity 40
2.3.4 Assessing SSC 41
2.3.4.1 Estimating 𝜌(Vq ) 41
2.3.4.2 Estimating the Significance of 𝜌(Vq ) 42
2.3.5 Empirical Studies 45
2.3.5.1 Datasets 45
2.3.5.2 Performance of DHT Approximation 45
2.3.5.3 Effectiveness on Synthetic Events 47
2.3.5.4 SSC of Real Event 49
2.3.5.5 Scalability of Sampling-alg 51
2.3.6 Discussions 51
2.4 Two-Event Structural Correlation 52
2.4.1 Preliminaries and Problem Formulation 52
2.4.2 Measuring TESC 53
2.4.2.1 The Test 54
2.4.2.2 Reference Nodes 56
2.4.3 Reference Node Sampling 56
2.4.3.1 Batch_BFS 57
2.4.3.2 Importance Sampling 58
2.4.3.3 Global Sampling in Whole Graph 61
2.4.3.4 Complexity Analysis 61
2.4.4 Experiments 62
2.4.4.1 Graph Datasets 62
2.4.4.2 Event Simulation Methodology 63
2.4.4.3 Performance Comparison 63
2.4.4.4 Batch Importance Sampling 65
2.4.4.5 Impact of Graph Density 66
2.4.4.6 Efficiency and Scalability 66
2.4.4.7 Real Events 68
2.4.5 Discussions 70
2.5 Conclusions 72
Acknowledgments 72
References 72

3 Spectral Graph Theory and Structural Analysis of Complex Networks:

An Introduction 75
Salissou Moutari and Ashraf Ahmed
3.1 Introduction 75
3.2 Graph Theory: Some Basic Concepts 76
3.2.1 Connectivity in Graphs 77
3.2.2 Subgraphs and Special Graphs 80
 Contents XI

3.3 Matrix Theory: Some Basic Concepts 81

3.3.1 Trace and Determinant of a Matrix 81
3.3.2 Eigenvalues and Eigenvectors of a Matrix 82
3.4 Graph Matrices 83
3.4.1 Adjacency Matrix 84
3.4.2 Incidence Matrix 84
3.4.3 Degree Matrix and Diffusion Matrix 85
3.4.4 Laplace Matrix 85
3.4.5 Cut-Set Matrix 86
3.4.6 Path Matrix 86
3.5 Spectral Graph Theory: Some Basic Results 86
3.5.1 Spectral Characterization of Graph Connectivity 87
3.5.1.1 Spectral Theory and Walks 88
3.5.2 Spectral Characteristics of some Special Graphs and Subgraphs 89
3.5.2.1 Tree 89
3.5.2.2 Bipartite Graph 89
3.5.2.3 Complete Graph 90
3.5.2.4 Regular Graph 90
3.5.2.5 Line Graph 90
3.5.3 Spectral Theory and Graph Colouring 91
3.5.4 Spectral Theory and Graph Drawing 91
3.6 Computational Challenges for Spectral Graph Analysis 91
3.6.1 Krylov Subspace Methods 91
3.6.2 Constrained Optimization Approach 94
3.7 Conclusion 94
References 95

4 Contagion in Interbank Networks 97

Grzegorz Hałaj and Christoffer Kok
4.1 Introduction 97
4.2 Research Context 99
4.3 Models 103
4.3.1 Simulated Networks 104
4.3.1.1 Probability Map 105
4.3.1.2 Interbank Network 105
4.3.1.3 Contagion Mechanism 107
4.3.1.4 Fire sales of Illiquid Portfolio 108
4.3.2 Systemic Probability Index 109
4.3.3 Endogenous Networks 110
4.3.3.1 Banks 113
4.3.3.2 First Round–Optimization of Interbank Assets 115
4.3.3.3 Second Round–Accepting Placements According to Funding
Needs 116
4.3.3.4 Third Round–Bargaining Game 117
4.3.3.5 Fourth Round–Price Adjustments 118
XII Contents

4.4 Results 119

4.4.1 Data 119
4.4.2 Simulated Networks 120
4.4.3 Structure of Endogenous Interbank Networks 123
4.5 Stress Testing Applications 127
4.6 Conclusions 130
References 131

5 Detection, Localization, and Tracking of a Single and Multiple Targets

with Wireless Sensor Networks 137
Natallia Katenka, Elizaveta Levina, and George Michailidis
5.1 Introduction and Overview 137
5.2 Data Collection and Fusion by WSN 138
5.3 Target Detection 141
5.3.1 Target Detection from Value Fusion (Energies) 142
5.3.2 Target Detection from Ordinary Decision Fusion 143
5.3.3 Target Detection from Local Vote Decision Fusion 144
5.3.3.1 Remark 1: LVDF Fixed Neighbourhood Size 145
5.3.3.2 Remark 2: LVDF Regular Grids 146
5.3.3.3 Remark 3: Quality of Approximation 148
5.3.3.4 Remark 4: Detection Performance 148
5.3.3.5 Concluding Remarks 148
5.4 Single Target Localization and Diagnostic 149
5.4.1 Localization and Diagnostic from Value Fusion (Energies) 150
5.4.2 Localization and Diagnostic from Ordinary Decision Fusion 151
5.4.3 Localization and Diagnostic from Local Vote Decision Fusion 152
5.4.4 Hybrid Maximum Likelihood Estimates 153
5.4.5 Properties of Maximum-Likelihood Estimates 154
5.4.5.1 Remark 1: Accuracy of Target Localization 155
5.4.5.2 Remark 2: Starting Values for Localization 155
5.4.5.3 Remark 3: Robustness to Model Misspecification 156
5.4.5.4 Remark 4: Computational Cost 156
5.4.5.5 Concluding Remarks 157
5.5 Multiple Target Localization and Diagnostic 157
5.5.1 Multiple Target Localization from Energies 158
5.5.2 Multiple Target Localization from Binary Decisions 158
5.5.3 Multiple Target Localization from Corrected Decisions 159
5.5.3.1 Remark 1: Hybrid Estimation 160
5.5.3.2 Remark 2: Starting Values 160
5.5.3.3 Estimating the Number of Targets 160
5.5.3.4 Concluding Remarks 160
5.6 Multiple Target Tracking 161
5.7 Applications and Case Studies 165
5.7.1 The NEST Project 166
5.7.2 The ZebraNet Project 168
 Contents XIII

5.8 Final Remarks 170

References 171

6 Computing in Dynamic Networks 173

Othon Michail, Ioannis Chatzigiannakis, and Paul G. Spirakis
6.1 Introduction 173
6.1.1 Motivation-State of the Art 173
6.1.2 Structure of the Chapter 177
6.2 Preliminaries 177
6.2.1 The Dynamic Network Model 177
6.2.2 Problem Definitions 179
6.3 Spread of Influence in Dynamic Graphs (Causal Influence) 180
6.4 Naming and Counting in Anonymous Unknown Dynamic
Networks 182
6.4.1 Further Related Work 183
6.4.2 Static Networks with Broadcast 183
6.4.3 Dynamic Networks with Broadcast 186
6.4.4 Dynamic Networks with One-to-Each 188
6.4.5 Higher Dynamicity 195
6.5 Causality, Influence, and Computation in Possibly Disconnected
Synchronous Dynamic Networks 196
6.5.1 Our Metrics 196
6.5.1.1 The Influence Time 196
6.5.1.2 The Moi (Concurrent Progress) 199
6.5.1.3 The Connectivity Time 200
6.5.2 Fast Propagation of Information under Continuous
Disconnectivity 201
6.5.3 Termination and Computation 203
6.5.3.1 Nodes Know an Upper Bound on the ct: An Optimal Termination
Criterion 204
6.5.3.2 Known Upper Bound on the oit 205
6.5.3.3 Hearing the Future 208
6.6 Local Communication Windows 212
6.7 Conclusions 215
References 216

7 Visualization and Interactive Analysis for Complex Networks by means

of Lossless Network Compression 219
Matthias Reimann, Loïc Royer, Simone Daminelli, and Michael Schroeder
7.1 Introduction 219
7.1.1 Illustrative Example 221
7.2 Power Graph Algorithm 221
7.2.1 Formal Definition of Power Graphs 221
7.2.2 Semantics of Power Graphs 222
7.2.3 Power Graph Conditions 222
XIV Contents

7.2.4 Edge Reduction and Relative Edge Reduction 223

7.2.5 Power Graph Extraction 225
7.3 Validation – Edge Reduction Differs from Random 227
7.4 Graph Comparison with Power Graphs 228
7.5 Excursus: Layout of Power Graphs 229
7.6 Interactive Visual Analytics 231
7.6.1 Power Edge Filtering 232
7.6.1.1 Zooming and Network Expansion 233
7.7 Conclusion 234
References 234

Index 237
 XV

Color Plates

DeVeaux Friedman
1.00

0.75

0.50

0.25
Percentage

0.00 n
300
Tibshirani Turlach 400
1.00 600

0.75

0.50

0.25

0.00
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
Size

Figure 1.9 Proportion of hidden layer size identification by using the testing procedure for
superior predictive ability. (This figure also appears on page 24.)
XVI Color Plates

0.95

0.9
Recall

0.85

0.8 Positive, h = 3, noise = 0.1

Positive, h = 2, noise = 0
0.75 Negative, h = 3, noise = 0
Negative, h = 2, noise = 0.5
0.7
5 10 15 20
# reference nodes sampled from Vvh

Figure 2.17 Performance of sampling different number of reference nodes from each Vvh
for Importance sampling. (This figure also appears on page 65.)

Figure 3.1 The graph G, drawn using its adjacency matrix A and randomly chosen
positions for the vertices. (This figure also appears on page 92.)
 Color Plates XVII

Figure 3.2 The graph G, drawn using its adjacency matrix A and its two Laplace eigenvec-
tors u2 (L) and u3 (L). (This figure also appears on page 92.)