Big Data and Visual Analytics (eBook)

Foreword 5
Contents 8
Automated Detection of Central Retinal Vein Occlusion Using Convolutional Neural Network 10
1 Introduction 11
2 Central Retinal Vein Occlusion (CRVO) 13
2.1 Non-Ischemic CRVO 14
2.2 Ischemic CRVO 14
3 Computer Aided Detection (CAD) 15
4 State of the Art for RVO Detection 17
4.1 Feature Representation Techniques 17
4.2 Fractal Analysis 17
4.3 Deep Learning Approach 18
5 The Proposed Methodology 19
5.1 The Basic of the Convolutional Neural Network (CNN) 20
5.1.1 The Convolutional Layer 21
5.1.2 The Pooling Layer 21
5.1.3 Non-Linear layer 21
5.1.4 Fully Connected Layers 21
5.2 Methodology 22
5.2.1 Image Preprocessing 22
5.2.2 The Network Topology 23
6 Result and Discussion 26
7 Conclusion 29
References 29
Swarm Intelligence Applied to Big Data Analytics for Rescue Operations with RASEN Sensor Networks 31
1 Introduction 32
2 Literature Survey 34
2.1 Rescue Drones 34
2.2 Drone Networking 36
2.3 Regional Disasters 40
2.4 Swarm Intelligence 44
2.5 Night Vision Systems 47
2.6 Artificial Cognitive Architectures 48
3 Rapid Alert System for Enhanced Night Vision (RASEN) 49
4 Swarm Intelligence Utilizing Networked RFID 53
4.1 Radio Frequency Identification (RFID) for Wireless Drone Networking 53
4.2 Ant-Colony Meta-Heuristics for Night Rescue Operations 56
5 Conclusion 58
References 59
Gender Classification Based on Deep Learning 63
1 Introduction 63
2 Related Works 64
3 Deep Learning 65
3.1 Background in Neural Networks 66
3.2 Convolutional Neural Networks 68
4 Data Sets 69
5 Deep Learning Architecture 70
5.1 Experimental Settings 71
5.2 Experimental Results 72
6 Conclusion 75
References 76
Social and Organizational Culture in Korea and Women's Career Development 78
1 Introduction 78
2 Korean Social Culture 79
2.1 Korea's Patriarchal, Male-Dominated Culture 80
2.2 Social Norms and the Socialization of Gender Roles 82
3 Korea's Economic Indicator and Current State of Women Labor 82
3.1 Korea's Productive Population and Its Competence 83
3.2 Korean Women's Qualitative Employment Indicators 83
3.3 Korean Women's Employment Stabilization Indicator 85
4 Female Management Staff and Female CEOs in Korea 87
4.1 Female Management Staff and Female CEOs in Korea 88
4.2 Positioning Gender Equality and Executive Ratio 88
5 Conclusion and Future Directions 89
References 91
Big Data Framework for Agile Business (BDFAB) As a Basis for Developing Holistic Strategies in Big Data Adoption 92
1 Introduction 92
2 Positioning the Need for Big Data and Agile Framework 93
3 Big Data and Agile Business: Literature Review 94
4 BDFAB: Overview of the Framework 96
5 BDFAB Modules (Building Blocks) and Their Interdisciplinary Mapping 98
5.1 Business Decisions 99
5.2 Data Science: Analytics, Context and Technologies 99
5.3 User Experience: Operational Excellence 100
5.4 Quality Dimensions and SMAC-Stack 100
5.5 People (Capability) 100
6 Conclusions and Future Direction 100
References 101
Scalable Gene Sequence Analysis on Spark 103
1 Introduction 103
2 Background 105
2.1 Apache Hadoop and Spark 105
2.2 Gene Sequence Analysis 106
3 System Model for Scalable Gene Sequence Analysis 107
4 Evaluation 108
4.1 Spark SQL vs. Other High-Level Query Tools (Pig and Hive) 109
4.2 Scale-Out and Scale-Up Performance 109
4.3 Impact of the Number of Cached Columns 112
4.4 Impact of Data Spills 112
4.5 Impact of Record Hits 113
5 System Prototyping 114
6 Conclusions 115
References 118
Big Sensor Data Acquisition and Archiving with Compression 120
1 Introduction 120
1.1 Sensor Data Acquisition 121
1.2 Sensor Data Archiving 122
2 Low Complexity Sampling 123
2.1 Compressive Sensing 123
2.1.1 General Signal Recovery 125
2.1.2 Noisy Signal Recovery 125
2.2 Random Sampling in Spatio-Temporal Dimension 126
2.2.1 Low Complexity Sampling Framework 127
2.2.2 Signal Recovery 129
2.3 Evaluation 130
3 Statistical Similarity Based Data Compression 133
3.1 Similarity Measure 133
3.2 IDEALEM Design 135
3.2.1 Encoded Stream Structure 135
3.2.2 Decoding 136
3.3 Evaluation 136
4 Scalable Management of Storage Space 138
4.1 Storage Space Optimization 138
4.2 Overview of Archiving Scheme 139
4.3 Gradual Decrease of Data Fidelity 140
4.4 Data Fidelity Model 141
4.5 Optimal Rate Allocation 143
4.6 Evaluation 143
5 Conclusion 144
References 145
Advanced High Performance Computing for Big Data Local Visual Meaning 149
1 Introduction 150
2 The Impact of Latency in Rapid Prototype Applications 151
3 Big Data on Analog Electronic Design Challenge 152
4 Missing Details in the Local Visual Meaning Analogue Works 158
5 Pilot Application: Adaptive BPF Analog Circuit 162
6 Conclusions and Future Directions 163
References 165
Transdisciplinary Benefits of Convergence in Big Data Analytics 169
1 Introduction 169
2 Background and Definition of Big Data 171
2.1 Big Data Origin 171
2.2 Big Data Definition 171
2.3 An Early Application of Big Data Approach 172
3 Disciplinary Examples of Big Data Applications 173
3.1 Big Data in the Healthcare Discipline 173
3.2 Big Data in the Energy Discipline 177
3.3 Big Data in the Business Discipline 179
4 Convergence: A Key Link Between Big Data, Platforms, and People 180
5 Conclusion 181
References 182
A Big Data Analytics Approach in Medical Imaging Segmentation Using Deep Convolutional Neural Networks 184
1 Introduction 184
2 Methods and Materials 186
2.1 Imaging Data 186
2.2 Brain Tumor Segmentation Method Based on Dense Convolutional Neural Network 187
2.3 Image Preprocessing 188
2.3.1 Dense Convolutional Neural Network Model 188
3 Experiment and Results 190
4 Conclusion 190
References 191
Big Data in Libraries 193
1 Introduction 193
2 Using Big Data In Libraries 194
2.1 Needs Analysis 195
2.2 Current Resources 195
2.3 Identify Solutions 196
2.4 Data Stores 196
2.5 Administration 197
2.6 Policies and Access 197
2.7 Analysis and Visualization 198
2.8 Query Access 198
2.9 Implementation 199
3 Library Big Data Services 199
3.1 Planning 199
3.2 Collection 200
3.3 Assure, Analyze and Integrate 200
3.4 Describe 201
3.5 Preserve 202
3.6 Discover 202
4 Conclusion 203
References 203
A Framework for Social Network Sentiment Analysis Using Big Data Analytics 205
1 Introduction 205
2 Research Questions of Thesis 206
2.1 Question 1: How Would It Be Possible to Parameterize Domain Characteristics in Order to Allow the Framework to Specialize for Various Domains? 206
2.2 Question 2: How Would It Be Possible to Create an Upgradable Framework Such that Components Can Be Replaced Whenever New Tools with Better Performance Become Available? 206
3 Domain Literature and Theoretical Background 207
3.1 Sentiment Analysis 207
3.1.1 Types of Sentiment Analysis 207
3.2 Machine Learning 207
3.2.1 Learning Process 208
3.2.2 Naive Bayes Algorithm 208
3.3 Big Data Analytics 210
4 System Architecture 210
5 Process Flow 211
5.1 Data Collection 211
5.2 Preprocessing 212
5.2.1 Language Detection 212
5.2.2 Lower Case 212
5.2.3 Regular Expressions 212
5.2.4 Tokenization 212
5.2.5 Parts of Speech Tagging 212
5.3 Data Modelling 213
6 Tools Profile 213
6.1 Jupyter Notebook 214
6.2 Anaconda Framework 214
6.3 Spark 214
7 Case Studies 214
7.1 Case Study I 214
7.2 Key Findings 216
7.3 Case Study II 217
8 Conclusions and Future Directions 218
References 218
Big Data Analytics and Visualization: Finance 220
1 Introduction 220
1.1 How Big Data Analytics Could benefit Finance Industry? 220
2 Areas Where Banking and Financial Institution Could Focus on Using Big Data and Analytics 222
3 Big Data with Machine Learning for Customer Centric Decision Making 222
4 Big Data for Sentiment Analysis Using Natural Language Programming 226
5 Big Data Analytics and Business Intelligence 227
6 Conclusion 228
References 230
Study of Hardware Trojans in a Closed Loop Control System for an Internet-of-Things Application 231
1 Introduction 231
2 Hardware Trojan Detection and Emerging Solutions 233
3 Design and Implementation 233
3.1 Discrete-Time Control System 233
3.2 Discrete PID Controller 234
3.3 Hardware Trojan Taxonomy 235
3.4 Threat Model Implementation in a Closed Loop Control System 236
3.5 Experimental Setup 237
3.6 Design Workflow 238
4 Results 239
4.1 Threat Model Implementation in PID Controller: Threat I—Turn Off Controller 239
4.2 Threat Model Implementation in PID Controller: Threat II—Turn Off Controller for a Short Time 240
4.3 Threat Model Implementation in PID Controller: Threat III: Turn Off and On Controller after 17 S 240
4.4 Threat Model Implementation in PID Controller: Threat IV—Delay in Controller: Variable Delay Length, and Threshold Delay vs No-Delay 241
5 Conclusion 243
References 243
High Performance/Throughput Computing Workflow for a Neuro-Imaging Application: Provenance and Approaches 244
1 Introduction 244
2 Methods 246
2.1 Typical Researcher Development Workflow 246
2.2 Typical Lab Development/Production Workflow 248
2.3 High Performance/High Throughput (HP/HT) Production Workflow (Local) 248
2.4 Improved HPC/HTC Workflow with Computing Support from a National Computing Facility 250
3 Test Setup, Assumptions and Results 252
3.1 Test Parameters 252
4 Discussion and Conclusion 252
References 254
Index 256