Machine Learning and Big Data

Uma N. Dulhare is a Professor in the Department of Computer Science & Eng., MJCET affiliated to Osmania University, Hyderabad, India. She has more than 20 years teaching experience years with many publications in reputed international conferences, journals and online book chapter contributions. She received her PhD from Osmania University, Hyderabad. Khaleel Ahmad is an Assistant Professor in the Department of Computer Science & Information Technology at Maulana Azad National Urdu University, Hyderabad, India. He holds a PhD in Computer Science & Engineering. He has published more than 25 papers in refereed journals and conferences as well as edited two books. Khairol Amali bin Ahmad obtained a BSc in Electrical Engineering in 1992 from the United States Military Academy, West Point, MSc in Military Electronic Systems Engineering in 1999 from Cranfield University, England, and PhD from ISAE-SUPAERO, France in 2015. Currently, he is the Dean of the Engineering Faculty at the National Defense University of Malaysia.

• Preface xixSection 1: Theoretical Fundamentals 11 Mathematical Foundation 3Afroz and Basharat Hussain1.1 Concept of Linear Algebra 31.1.1 Introduction 31.1.2 Vector Spaces 51.1.3 Linear Combination 61.1.4 Linearly Dependent and Independent Vectors 71.1.5 Linear Span, Basis and Subspace 81.1.6 Linear Transformation (or Linear Map) 91.1.7 Matrix Representation of Linear Transformation 101.1.8 Range and Null Space of Linear Transformation 131.1.9 Invertible Linear Transformation 151.2 Eigenvalues, Eigenvectors, and Eigendecomposition of a Matrix 151.2.1 Characteristics Polynomial 161.2.1.1 Some Results on Eigenvalue 161.2.2 Eigendecomposition 181.3 Introduction to Calculus 201.3.1 Function 201.3.2 Limits of Functions 211.3.2.1 Some Properties of Limits 221.3.2.2 1nfinite Limits 251.3.2.3 Limits at Infinity 261.3.3 Continuous Functions and Discontinuous Functions 261.3.3.1 Discontinuous Functions 271.3.3.2 Properties of Continuous Function 271.3.4 Differentiation 28References 292 Theory of Probability 31Parvaze Ahmad Dar and Afroz2.1 Introduction 312.1.1 Definition 312.1.1.1 Statistical Definition of Probability 312.1.1.2 Mathematical Definition of Probability 322.1.2 Some Basic Terms of Probability 322.1.2.1 Trial and Event 322.1.2.2 Exhaustive Events (Exhaustive Cases) 332.1.2.3 Mutually Exclusive Events 332.1.2.4 Equally Likely Events 332.1.2.5 Certain Event or Sure Event 332.1.2.6 Impossible Event or Null Event (ϕ) 332.1.2.7 Sample Space 342.1.2.8 Permutation and Combination 342.1.2.9 Examples 352.2 Independence in Probability 382.2.1 Independent Events 382.2.2 Examples: Solve the Following Problems 382.3 Conditional Probability 412.3.1 Definition 412.3.2 Mutually Independent Events 422.3.3 Examples 422.4 Cumulative Distribution Function 432.4.1 Properties 442.4.2 Example 442.5 Baye’s Theorem 462.5.1 Theorem 462.5.1.1 Examples 472.6 Multivariate Gaussian Function 502.6.1 Definition 502.6.1.1 Univariate Gaussian (i.e., One Variable Gaussian) 502.6.1.2 Degenerate Univariate Gaussian 512.6.1.3 Multivariate Gaussian 51References 513 Correlation and Regression 53Mohd. Abdul Haleem Rizwan3.1 Introduction 533.2 Correlation 543.2.1 Positive Correlation and Negative Correlation 543.2.2 Simple Correlation and Multiple Correlation 543.2.3 Partial Correlation and Total Correlation 543.2.4 Correlation Coefficient 553.3 Regression 573.3.1 Linear Regression 643.3.2 Logistic Regression 643.3.3 Polynomial Regression 653.3.4 Stepwise Regression 663.3.5 Ridge Regression 673.3.6 Lasso Regression 673.3.7 Elastic Net Regression 683.4 Conclusion 68References 69Section 2: Big Data and Pattern Recognition 714 Data Preprocess 73Md. Sharif Hossen4.1 Introduction 734.1.1 Need of Data Preprocessing 744.1.2 Main Tasks in Data Preprocessing 754.2 Data Cleaning 774.2.1 Missing Data 774.2.2 Noisy Data 784.3 Data Integration 804.3.1 χ2 Correlation Test 824.3.2 Correlation Coefficient Test 824.3.3 Covariance Test 834.4 Data Transformation 834.4.1 Normalization 834.4.2 Attribute Selection 854.4.3 Discretization 864.4.4 Concept Hierarchy Generation 864.5 Data Reduction 884.5.1 Data Cube Aggregation 884.5.2 Attribute Subset Selection 904.5.3 Numerosity Reduction 914.5.4 Dimensionality Reduction 954.6 Conclusion 101Acknowledgements 101References 1015 Big Data 105R. Chinnaiyan5.1 Introduction 1055.2 Big Data Evaluation With Its Tools 1075.3 Architecture of Big Data 1075.3.1 Big Data Analytics Framework Workflow 1075.4 Issues and Challenges 1095.4.1 Volume 1095.4.2 Variety of Data 1105.4.3 Velocity 1105.5 Big Data Analytics Tools 1105.6 Big Data Use Cases 1145.6.1 Banking and Finance 1145.6.2 Fraud Detection 1145.6.3 Customer Division and Personalized Marketing 1145.6.4 Customer Support 1155.6.5 Risk Management 1165.6.6 Life Time Value Prediction 1165.6.7 Cyber Security Analytics 1175.6.8 Insurance Industry 1185.6.9 Health Care Sector 1185.6.9.1 Big Data Medical Decision Support 1205.6.9.2 Big Data–Based Disorder Management 1205.6.9.3 Big Data–Based Patient Monitoring and Control 1205.6.9.4 Big Data–Based Human Routine Analytics 1205.6.10 Internet of Things 1215.6.11 Weather Forecasting 1215.7 Where IoT Meets Big Data 1225.7.1 IoT Platform 1225.7.2 Sensors or Devices 1235.7.3 Device Aggregators 1235.7.4 IoT Gateway 1235.7.5 Big Data Platform and Tools 1245.8 Role of Machine Learning For Big Data and IoT 1245.8.1 Typical Machine Learning Use Cases 1255.9 Conclusion 126References 1276 Pattern Recognition Concepts 131Ambeshwar Kumar, R. Manikandan and C. Thaventhiran6.1 Classifier 1326.1.1 Introduction 1326.1.2 Explanation-Based Learning 1336.1.3 Isomorphism and Clique Method 1356.1.4 Context-Dependent Classification 1386.1.5 Summary 1396.2 Feature Processing 1406.2.1 Introduction 1406.2.2 Detection and Extracting Edge With Boundary Line 1416.2.3 Analyzing the Texture 1426.2.4 Feature Mapping in Consecutive Moving Frame 1436.2.5 Summary 1456.3 Clustering 1456.3.1 Introduction 1456.3.2 Types of Clustering Algorithms 1466.3.2.1 Dynamic Clustering Method 1486.3.2.2 Model-Based Clustering 1486.3.3 Application 1496.3.4 Summary 1506.4 Conclusion 151References 151Section 3: Machine Learning: Algorithms & Applications 1537 Machine Learning 155Elham Ghanbari and Sara Najafzadeh7.1 History and Purpose of Machine Learning 1557.1.1 History of Machine Learning 1557.1.1.1 What is Machine Learning? 1567.1.1.2 When the Machine Learning is Needed? 1577.1.2 Goals and Achievements in Machine Learning 1587.1.3 Applications of Machine Learning 1587.1.3.1 Practical Machine Learning Examples 1597.1.4 Relation to Other Fields 1617.1.4.1 Data Mining 1617.1.4.2 Artificial Intelligence 1627.1.4.3 Computational Statistics 1627.1.4.4 Probability 1637.1.5 Limitations of Machine Learning 1637.2 Concept of Well-Defined Learning Problem 1647.2.1 Concept Learning 1647.2.1.1 Concept Representation 1667.2.1.2 Instance Representation 1677.2.1.3 The Inductive Learning Hypothesis 1677.2.2 Concept Learning as Search 1677.2.2.1 Concept Generality 1687.3 General-to-Specific Ordering Over Hypotheses 1697.3.1 Basic Concepts: Hypothesis, Generality 1697.3.2 Structure of the Hypothesis Space 1697.3.2.1 Hypothesis Notations 1697.3.2.2 Hypothesis Evaluations 1707.3.3 Ordering on Hypotheses: General to Specific 1707.3.3.1 Most Specific Generalized 1717.3.3.2 Most General Specialized 1737.3.3.3 Generalization and Specialization Operators 1737.3.4 Hypothesis Space Search by Find-S Algorithm 1747.3.4.1 Properties of the Find-S Algorithm 1767.3.4.2 Limitations of the Find-S Algorithm 1767.4 Version Spaces and Candidate Elimination Algorithm 1777.4.1 Representing Version Spaces 1777.4.1.1 General Boundary 1787.4.1.2 Specific Boundary 1787.4.2 Version Space as Search Strategy 1797.4.3 The List-Eliminate Method 1797.4.4 The Candidate-Elimination Method 1807.4.4.1 Example 1817.4.4.2 Convergence of Candidate-Elimination Method 1837.4.4.3 Inductive Bias for Candidate-Elimination 1847.5 Concepts of Machine Learning Algorithm 1857.5.1 Types of Learning Algorithms 1857.5.1.1 Incremental vs. Batch Learning Algorithms 1867.5.1.2 Offline vs. Online Learning Algorithms 1887.5.1.3 Inductive vs. Deductive Learning Algorithms 1897.5.2 A Framework for Machine Learning Algorithms 1897.5.2.1 Training Data 1907.5.2.2 Target Function 1907.5.2.3 Construction Model 1917.5.2.4 Evaluation 1917.5.3 Types of Machine Learning Algorithms 1947.5.3.1 Supervised Learning 1967.5.3.2 Unsupervised Learning 1987.5.3.3 Semi Supervised Learning 2007.5.3.4 Reinforcement Learning 2007.5.3.5 Deep Learning 2027.5.4 Types of Machine Learning Problems 2037.5.4.1 Classification 2047.5.4.2 Clustering 2047.5.4.3 Optimization 2057.5.4.4 Regression 205Conclusion 205References 2068 Performance of Supervised Learning Algorithms on Multi-Variate Datasets 209Asif Iqbal Hajamydeen and Rabab Alayham Abbas Helmi8.1 Introduction 2098.2 Supervised Learning Algorithms 2108.2.1 Datasets and Experimental Setup 2118.2.2 Data Treatment/Preprocessing 2128.3 Classification 2128.3.1 Support Vector Machines (SVM) 2138.3.2 Naive Bayes (NB) Algorithm 2148.3.3 Bayesian Network (BN) 2148.3.4 Hidden Markov Model (HMM) 2158.3.5 K-Nearest Neighbour (KNN) 2168.3.6 Training Time 2168.4 Neural Network 2178.4.1 Artificial Neural Networks Architecture 2198.4.2 Application Areas 2228.4.3 Artificial Neural Networks and Time Series 2248.5 Comparisons and Discussions 2258.5.1 Comparison of Classification Accuracy 2258.5.2 Forecasting Efficiency Comparison 2268.5.3 Recurrent Neural Network (RNN) 2268.5.4 Backpropagation Neural Network (BPNN) 2288.5.5 General Regression Neural Network 2298.6 Summary and Conclusion 230References 2319 Unsupervised Learning 233M. Kumara Swamy and Tejaswi Puligilla9.1 Introduction 2339.2 Related Work 2349.3 Unsupervised Learning Algorithms 2359.4 Classification of Unsupervised Learning Algorithms 2389.4.1 Hierarchical Methods 2389.4.2 Partitioning Methods 2399.4.3 Density-Based Methods 2429.4.4 Grid-Based Methods 2459.4.5 Constraint-Based Clustering 2459.5 Unsupervised Learning Algorithms in ML 2469.5.1 Parametric Algorithms 2469.5.2 Non-Parametric Algorithms 2469.5.3 Dirichlet Process Mixture Model 2479.5.4 X-Means 2489.6 Summary and Conclusions 248References 24810 Semi-Supervised Learning 251Manish Devgan, Gaurav Malik and Deepak Kumar Sharma10.1 Introduction 25210.1.1 Semi-Supervised Learning 25210.1.2 Comparison With Other Paradigms 25510.2 Training Models 25710.2.1 Self-Training 25710.2.2 Co-Training 25910.3 Generative Models—Introduction 26110.3.1 Image Classification 26410.3.2 Text Categorization 26610.3.3 Speech Recognition 26810.3.4 Baum-Welch Algorithm 26810.4 S3VMs 27010.5 Graph-Based Algorithms 27410.5.1 Mincut 27510.5.2 Harmonic 27610.5.3 Manifold Regularization 27710.6 Multiview Learning 27710.7 Conclusion 278References 27911 Reinforcement Learning 281Amandeep Singh Bhatia, Mandeep Kaur Saggi, Amit Sundas and Jatinder Ashta11.1 Introduction: Reinforcement Learning 28111.1.1 Elements of Reinforcement Learning 28311.2 Model-Free RL 28411.2.1 Q-Learning 28511.2.2 R-Learning 28611.3 Model-Based RL 28711.3.1 SARSA Learning 28911.3.2 Dyna-Q Learning 29011.3.3 Temporal Difference 29111.3.3.1 TD(0) Algorithm 29211.3.3.2 TD(1) Algorithm 29311.3.3.3 TD(λ) Algorithm 29411.3.4 Monte Carlo Method 29411.3.4.1 Monte Carlo Reinforcement Learning 29611.3.4.2 Monte Carlo Policy Evaluation 29611.3.4.3 Monte Carlo Policy Improvement 29811.4 Conclusion 298References 29912 Application of Big Data and Machine Learning 305Neha Sharma, Sunil Kumar Gautam, Azriel A. Henry and Abhimanyu Kumar12.1 Introduction 30612.2 Motivation 30712.3 Related Work 30812.4 Application of Big Data and ML 30912.4.1 Healthcare 30912.4.2 Banking and Insurance 31212.4.3 Transportation 31412.4.4 Media and Entertainment 31612.4.5 Education 31712.4.6 Ecosystem Conservation 31912.4.7 Manufacturing 32112.4.8 Agriculture 32212.5 Issues and Challenges 32412.6 Conclusion 326References 326Section 4: Machine Learning’s Next Frontier 33513 Transfer Learning 337Riyanshi Gupta, Kartik Krishna Bhardwaj and Deepak Kumar Sharma13.1 Introduction 33813.1.1 Motivation, Definition, and Representation 33813.2 Traditional Learning vs. Transfer Learning 33813.3 Key Takeaways: Functionality 34013.4 Transfer Learning Methodologies 34113.5 Inductive Transfer Learning 34213.6 Unsupervised Transfer Learning 34413.7 Transductive Transfer Learning 34613.8 Categories in Transfer Learning 34713.9 Instance Transfer 34813.10 Feature Representation Transfer 34913.11 Parameter Transfer 34913.12 Relational Knowledge Transfer 35013.13 Relationship With Deep Learning 35113.13.1 Transfer Learning in Deep Learning 35113.13.2 Types of Deep Transfer Learning 35213.13.3 Adaptation of Domain 35213.13.4 Domain Confusion 35313.13.5 Multitask Learning 35413.13.6 One-Shot Learning 35413.13.7 Zero-Shot Learning 35513.14 Applications: Allied Classical Problems 35513.14.1 Transfer Learning for Natural Language Processing 35613.14.2 Transfer Learning for Computer Vision 35613.14.3 Transfer Learning for Audio and Speech 35713.15 Further Advancements and Conclusion 357References 358Section 5: Hands-On and Case Study 36114 Hands on MAHOUT—Machine Learning ToolUma N. Dulhare and Sheikh Gouse14.1 Introduction to Mahout 36314.1.1 Features 36614.1.2 Advantages 36614.1.3 Disadvantages 36614.1.4 Application 36614.2 Installation Steps of Apache Mahout Using Cloudera 36714.2.1 Installation of VMware Workstation 36714.2.2 Installation of Cloudera 36814.2.3 Installation of Mahout 38314.2.4 Installation of Maven 38414.2.5 Testing Mahout 38614.3 Installation Steps of Apache Mahout Using Windows 10 38614.3.1 Installation of Java 38614.3.2 Installation of Hadoop 38714.3.3 Installation of Mahout 38714.3.4 Installation of Maven 38714.3.5 Path Setting 38814.3.6 Hadoop Configuration 39114.4 Installation Steps of Apache Mahout Using Eclipse 39514.4.1 Eclipse Installation 39514.4.2 Installation of Maven Through Eclipse 39614.4.3 Maven Setup for Mahout Configuration 39914.4.4 Building the Path- 40214.4.5 Modifying the pom.xml File 40514.4.6 Creating the Data File 40714.4.7 Adding External Jar Files 40814.4.8 Creating the New Package and Classes 41014.4.9 Result 41114.5 Mahout Algorithms 41214.5.1 Classification 41214.5.2 Clustering 41314.5.3 Recommendation 41514.6 Conclusion 418References 41815 Hands-On H2O Machine Learning Tool 423Uma N. Dulhare, Azmath Mubeen and Khaleel Ahmed15.1 Introduction 42415.2 Installation 42515.2.1 The Process of Installation 42515.3 Interfaces 43115.4 Programming Fundamentals 43215.4.1 Data Manipulation 43215.4.1.1 Data Types 43215.4.1.2 Data Import 43515.4.2 Models 43615.4.2.1 Model Training 43615.4.3 Discovering Aspects 43715.4.3.1 Converting Data Frames 43715.4.4 H2O Cluster Actions 43815.4.4.1 H2O Key Value Retrieval 43815.4.4.2 H2O Cluster Connection 43815.4.5 Commands 43915.4.5.1 Cluster Information 43915.4.5.2 General Data Operations 44115.4.5.3 String Manipulation Commands 44215.5 Machine Learning in H2O 44215.5.1 Supervised Learning 44215.5.2 Unsupervised Learning 44315.6 Applications of H2O 44315.6.1 Deep Learning 44315.6.2 K-Fold Cross-Authentication or Validation 44815.6.3 Stacked Ensemble and Random Forest Estimator 45015.7 Conclusion 452References 45316 Case Study: Intrusion Detection System Using Machine Learning 455Syeda Hajra Mahin, Fahmina Taranum and Reshma Nikhat16.1 Introduction 45616.1.1 Components Used to Design the Scenario Include 45616.1.1.1 Black Hole 45616.1.1.2 Intrusion Detection System 45716.1.1.3 Components Used From MATLAB Simulator 45816.2 System Design 46516.2.1 Three Sub-Network Architecture 46516.2.2 Using Classifiers of MATLAB 46516.3 Existing Proposals 46716.4 Approaches Used in Designing the Scenario 46916.4.1 Algorithm Used in QualNet 46916.4.2 Algorithm Applied in MATLAB 47116.5 Result Analysis 47116.5.1 Results From QualNet 47116.5.1.1 Deployment 47116.5.1.2 Detection 47216.5.1.3 Avoidance 47316.5.1.4 Validation of Conclusion 47316.5.2 Applying Results to MATLAB 47316.5.2.1 K-Nearest Neighbor 47516.5.2.2 SVM 47716.5.2.3 Decision Tree 47716.5.2.4 Naive Bayes 47916.5.2.5 Neural Network 47916.6 Conclusion 484References 48417 Inclusion of Security Features for Implications of Electronic Governance Activities 487Prabal Pratap and Nripendra Dwivedi17.1 Introduction 48717.2 Objective of E-Governance 49117.3 Role of Identity in E-Governance 49317.3.1 Identity 49317.3.2 Identity Management and its Buoyancy Against Identity Theft in E-Governance 49417.4 Status of E-Governance in Other Countries 49617.4.1 E-Governance Services in Other Countries Like Australia and South Africa 49617.4.2 Adaptation of Processes and Methodology for Developing Countries 49617.4.3 Different Programs Related to E-Governance 49917.5 Pros and Cons of E-Governance 50117.6 Challenges of E-Governance in Machine Learning 50217.7 Conclusion 503References 503Index 505