Big-Data Analytics for Cloud, IoT and Cognitive Computing

Kai Hwang, PhD is Professor of Electrical Engineering and Computer Science at University of Southern California, USA. He also serves as an EMC-endowed visiting Chair Professor at Tsinghua University, China. He specializes in computer architecture, wireless Internet, cloud computing and network security. Min Chen, PhD is Professor of Computer Science and Technology, Huazhong University of Science and Technology, China. His work focuses on IoT, mobile cloud, body area networks, healthcare big-data and cyber physical systems.

• About the Authors xiPreface xiiiAbout the Companion Website xviiPart 1 Big Data, Clouds and Internet of Things 11 Big Data Science and Machine Intelligence 31.1 Enabling Technologies for Big Data Computing 31.1.1 Data Science and Related Disciplines 41.1.2 Emerging Technologies in the Next Decade 71.1.3 Interactive SMACT Technologies 131.2 Social-Media, Mobile Networks and Cloud Computing 161.2.1 Social Networks and Web Service Sites 171.2.2 Mobile Cellular Core Networks 191.2.3 Mobile Devices and Internet Edge Networks 201.2.4 Mobile Cloud Computing Infrastructure 231.3 Big Data Acquisition and Analytics Evolution 241.3.1 Big Data Value Chain Extracted from Massive Data 241.3.2 Data Quality Control, Representation and Database Models 261.3.3 Big Data Acquisition and Preprocessing 271.3.4 Evolving Data Analytics over the Clouds 301.4 Machine Intelligence and Big Data Applications 321.4.1 Data Mining and Machine Learning 321.4.2 Big Data Applications – An Overview 341.4.3 Cognitive Computing – An Introduction 381.5 Conclusions 42Homework Problems 42References 432 Smart Clouds, Virtualization and Mashup Services 452.1 Cloud Computing Models and Services 452.1.1 Cloud Taxonomy based on Services Provided 462.1.2 Layered Development Cloud Service Platforms 502.1.3 Cloud Models for Big Data Storage and Processing 522.1.4 Cloud Resources for Supporting Big Data Analytics 552.2 Creation of Virtual Machines and Docker Containers 572.2.1 Virtualization of Machine Resources 582.2.2 Hypervisors and Virtual Machines 602.2.3 Docker Engine and Application Containers 622.2.4 Deployment Opportunity of VMs/Containers 642.3 Cloud Architectures and Resources Management 652.3.1 Cloud Platform Architectures 652.3.2 VM Management and Disaster Recovery 682.3.3 OpenStack for Constructing Private Clouds 702.3.4 Container Scheduling and Orchestration 742.3.5 VMWare Packages for Building Hybrid Clouds 752.4 Case Studies of IaaS, PaaS and SaaS Clouds 772.4.1 AWS Architecture over Distributed Datacenters 782.4.2 AWS Cloud Service Offerings 792.4.3 Platform PaaS Clouds – Google AppEngine 832.4.4 Application SaaS Clouds – The Salesforce Clouds 862.5 Mobile Clouds and Inter-Cloud Mashup Services 882.5.1 Mobile Clouds and Cloudlet Gateways 882.5.2 Multi-Cloud Mashup Services 912.5.3 Skyline Discovery of Mashup Services 952.5.4 Dynamic Composition of Mashup Services 962.6 Conclusions 98Homework Problems 98References 1033 IoT Sensing, Mobile and Cognitive Systems 1053.1 Sensing Technologies for Internet of Things 1053.1.1 Enabling Technologies and Evolution of IoT 1063.1.2 Introducing RFID and Sensor Technologies 1083.1.3 IoT Architectural and Wireless Support 1103.2 IoT Interactions with GPS, Clouds and Smart Machines 1113.2.1 Local versus Global Positioning Technologies 1113.2.2 Standalone versus Cloud-Centric IoT Applications 1143.2.3 IoT Interaction Frameworks with Environments 1163.3 Radio Frequency Identification (RFID) 1193.3.1 RFID Technology and Tagging Devices 1193.3.2 RFID System Architecture 1203.3.3 IoT Support of Supply Chain Management 1223.4 Sensors, Wireless Sensor Networks and GPS Systems 1243.4.1 Sensor Hardware and Operating Systems 1243.4.2 Sensing through Smart Phones 1303.4.3 Wireless Sensor Networks and Body Area Networks 1313.4.4 Global Positioning Systems 1343.5 Cognitive Computing Technologies and Prototype Systems 1393.5.1 Cognitive Science and Neuroinformatics 1393.5.2 Brain-Inspired Computing Chips and Systems 1403.5.3 Google’s Brain Team Projects 1423.5.4 IoT Contexts for Cognitive Services 1453.5.5 Augmented and Virtual Reality Applications 1463.6 Conclusions 149Homework Problems 150References 152Part 2 Machine Learning and Deep Learning Algorithms 1554 Supervised Machine Learning Algorithms 1574.1 Taxonomy of Machine Learning Algorithms 1574.1.1 Machine Learning Based on Learning Styles 1584.1.2 Machine Learning Based on Similarity Testing 1594.1.3 Supervised Machine Learning Algorithms 1624.1.4 Unsupervised Machine Learning Algorithms 1634.2 Regression Methods for Machine Learning 1644.2.1 Basic Concepts of Regression Analysis 1644.2.2 Linear Regression for Prediction and Forecast 1664.2.3 Logistic Regression for Classification 1694.3 Supervised Classification Methods 1714.3.1 Decision Trees for Machine Learning 1714.3.2 Rule-based Classification 1754.3.3 The Nearest Neighbor Classifier 1814.3.4 Support Vector Machines 1834.4 Bayesian Network and Ensemble Methods 1874.4.1 Bayesian Classifiers 1884.4.2 Bayesian Belief Networks 1914.4.3 Random Forests and Ensemble Methods 1954.5 Conclusions 200Homework Problems 200References 2035 Unsupervised Machine Learning Algorithms 2055.1 Introduction and Association Analysis 2055.1.1 Introduction to Unsupervised Machine Learning 2055.1.2 Association Analysis and A priori Principle 2065.1.3 Association Rule Generation 2105.2 Clustering Methods without Labels 2135.2.1 Cluster Analysis for Prediction and Forecasting 2135.2.2 K-means Clustering for Classification 2145.2.3 Agglomerative Hierarchical Clustering 2175.2.4 Density-based Clustering 2215.3 Dimensionality Reduction and Other Algorithms 2255.3.1 Dimensionality Reduction Methods 2255.3.2 Principal Component Analysis (PCA) 2265.3.3 Semi-Supervised Machine Learning Methods 2315.4 How to Choose Machine Learning Algorithms? 2335.4.1 Performance Metrics and Model Fitting 2335.4.2 Methods to Reduce Model Over-Fitting 2375.4.3 Methods to Avoid Model Under-Fitting 2405.4.4 Effects of Using Different Loss Functions 2425.5 Conclusions 243Homework Problems 243References 2476 Deep Learning with Artificial Neural Networks 2496.1 Introduction 2496.1.1 Deep Learning Mimics Human Senses 2496.1.2 Biological Neurons versus Artificial Neurons 2516.1.3 Deep Learning versus Shallow Learning 2546.2 Artificial Neural Networks (ANN) 2566.2.1 Single Layer Artificial Neural Networks 2566.2.2 Multilayer Artificial Neural Network 2576.2.3 Forward Propagation and Back Propagation in ANN 2586.3 Stacked AutoEncoder and Deep Belief Network 2646.3.1 AutoEncoder 2646.3.2 Stacked AutoEncoder 2676.3.3 Restricted Boltzmann Machine 2696.3.4 Deep Belief Networks 2756.4 Convolutional Neural Networks (CNN) and Extensions 2776.4.1 Convolution in CNN 2776.4.2 Pooling in CNN 2806.4.3 Deep Convolutional Neural Networks 2826.4.4 Other Deep Learning Networks 2836.5 Conclusions 287Homework Problems 288References 291Part 3 Big Data Analytics for Health-Care and Cognitive Learning 2937 Machine Learning for Big Data in Healthcare Applications 2957.1 Healthcare Problems and Machine Learning Tools 2957.1.1 Healthcare and Chronic Disease Detection Problem 2957.1.2 Software Libraries for Machine Learning Applications 2987.2 IoT-based Healthcare Systems and Applications 2997.2.1 IoT Sensing for Body Signals 3007.2.2 Healthcare Monitoring System 3017.2.3 Physical Exercise Promotion and Smart Clothing 3047.2.4 Healthcare Robotics and Mobile Health Cloud 3057.3 Big Data Analytics for Healthcare Applications 3107.3.1 Healthcare Big Data Preprocessing 3107.3.2 Predictive Analytics for Disease Detection 3127.3.3 Performance Analysis of Five Disease Detection Methods 3167.3.4 Mobile Big Data for Disease Control 3207.4 Emotion-Control Healthcare Applications 3227.4.1 Mental Healthcare System 3237.4.2 Emotion-Control Computing and Services 3237.4.3 Emotion Interaction through IoT and Clouds 3277.4.4 Emotion-Control via Robotics Technologies 3297.4.5 A 5G Cloud-Centric Healthcare System 3327.5 Conclusions 335Homework Problems 336References 3398 Deep Reinforcement Learning and Social Media Analytics 3438.1 Deep Learning Systems and Social Media Industry 3438.1.1 Deep Learning Systems and Software Support 3438.1.2 Reinforcement Learning Principles 3468.1.3 Social-Media Industry and Global Impact 3478.2 Text and Image Recognition using ANN and CNN 3488.2.1 Numeral Recognition using TensorFlow for ANN 3498.2.2 Numeral Recognition using Convolutional Neural Networks 3528.2.3 Convolutional Neural Networks for Face Recognition 3568.2.4 Medical Text Analytics by Convolutional Neural Networks 3578.3 DeepMind with Deep Reinforcement Learning 3628.3.1 Google DeepMind AI Programs 3628.3.2 Deep Reinforcement Learning Algorithm 3648.3.3 Google AlphaGo Game Competition 3678.3.4 Flappybird Game using Reinforcement Learning 3718.4 Data Analytics for Social-Media Applications 3758.4.1 Big Data Requirements in Social-Media Applications 3758.4.2 Social Networks and Graph Analytics 3778.4.3 Predictive Analytics Software Tools 3838.4.4 Community Detection in Social Networks 3868.5 Conclusions 390Homework Problems 391References 393Index 395