Principles of Data Science (Transactions on Computational Science and Computational Intelligence)

Preface
Acknowledgments
Contents
About the Editors
Simulation-Based Data Acquisition
	1 Introduction
	2 What Is Computer Simulation?
	3 Computer Simulation for the Acquisition of Data
	4 Design and Execution of Experiments
	5 Simulation Frameworks and Toolkits
	6 Investigation of a Credit Market
	7 Conclusions
	References
Coding of Bits for Entities by Means of Discrete Events (CBEDE): A Method of Compression and Transmission of Data
	1 Introduction
	2 Data Science
	3 Traditional Technologies for Telecommunications
	4 Proposed Method: CBEDE
	5 Results and Discussion
	6 Conclusions
	References
Big Biomedical Data Engineering
	1 Introduction
	2 Big Data
	3 Big Data Analytics
		3.1 Descriptive Analytics
		3.2 Predictive Analytics
		3.3 Decision Analytics
		3.4 Prescriptive Analytics
		3.5 Security Analytics
	4 Applications of Big Data Analytics in Medical Research
		4.1 Neurology Research
		4.2 Genome Research
		4.3 Cancer Research
		4.4 Healthcare
	5 Issues and Challenges
		5.1 Data Gathering
		5.2 Data Storage
		5.3 Data Processing and Visualization
		5.4 Value
		5.5 Academic Research
		5.6 Data Privacy
		5.7 Real-Time Processing
	6 Opportunity
	7 Biomedical Future
	8 Conclusions
	References
Big Data Preprocessing: An Application on Online Social Networks
	1 Introduction
	2 Data Quality Issues
		2.1 Data Context
		2.2 Data Analysis Types
		2.3 Issues Solved by Preprocessing Tasks
		2.4 Text Processing
		2.5 Entity Resolution
		2.6 Data Extraction
	3 Big Data Cleansing Approaches
		3.1 Error Identification
		3.2 Dependency Rules
		3.3 Statistics and Machine Learning
	4 Machine Learning Algorithms for Preprocessing and Analysis
		4.1 Supervised Machine Learning
			4.1.1 Imbalanced Data
			4.1.2 Algorithms and Preprocessing Requirements
		4.2 Unsupervised Machine Learning
			4.2.1 Imputation of Missing Values
			4.2.2 Anomaly Detection
		4.3 Extracting and Integrating Data
			4.3.1 Data Warehouses
	5 Feature Engineering and Text Preparation
		5.1 Transformation of Features or Samples
		5.2 Dimensionality Reduction
		5.3 Natural Language Processing
			5.3.1 Text Extraction
			5.3.2 Feature Selection and Extraction
	6 Preprocessing Frameworks
		6.1 Distributed Frameworks
		6.2 Preprocessing Frameworks for OSN
	7 Conclusion
	References
Feature Engineering
	1 Introduction: Basic Concepts
	2 Feature Design
	3 Advanced Feature Space Transformations for Dimensionality Reduction and Data Space Optimization
		3.1 Feature Extraction and Dimensionality Reduction for Feature Space Optimization
		3.2 PCA (Principal Component Analysis)
			3.2.1 Unsupervised PCA Common Version
			3.2.2 Supervised PCA
			3.2.3 Kernel PCA (the Nonlinear Case)
			3.2.4 Observations About the PCA Usage
		3.3 LDA (Linear Discriminant Analysis)
	4 Feature Selection for Dimensionality Reduction
		4.1 Searching Algorithms for Feature Selection
		4.2 Evaluation Criteria for Feature Selection
	References
Data Summarization Using Sampling Algorithms: Data Stream Case Study
	1 Introduction
	2 Data Stream Basic Concepts
		2.1 Definition
		2.2 Data Stream Structure
		2.3 Time Modeling and Windowing Models
	3 Data Stream Application Domains
		3.1 Sensor Networks
		3.2 Financial Analysis
		3.3 Network Traffic Analysis
	4 Data Stream Management
	5 Data Summarization Using Sampling Algorithms
	6 Performance Evaluation of Sampling Algorithms
	7 Future Directions
		7.1 Sampling Impact on Data Stream Statistical Inference
		7.2 Adaptive Sampling
	8 Conclusion
	References
Fast Imputation: An Algorithmic Formalism
	1 Introduction
	2 Literature Review
	3 Headway in Statistical Imputation Approaches
	4 Requisites for Imputation Reliability
		4.1 Property-1. Imputation Fairness
		4.2 Property-2. Composition Ratio Should Be Relative Ratio of Observed Set
	5 Comprehensive Analysis of Inferential Algorithms
	6 Algorithmic Formalism
	7 Problem Modelling
		7.1 Score Prediction
	8 Missing Value Imputation Using Similar Cluster Analysis
	9 Problem Statement
		9.1 Objective
		9.2 Part A: Data Engineering
		9.3 Part B: Feature Engineering
		9.4 Part C: State-of-the-Art Imputation Techniques
		9.5 Part D: Filling Strategy
		9.6 Part E: Missing Data Recovery
		9.7 Part F: Mitigation of Missing Values
	10 Methodology
	11 Approximation
	12 Proposed Algorithm
	13 Simulation Results
	14 Discussion and Future Works
	A.1 Appendix: Review of State-of-the-Art [41–46] Imputation Algorithms
	References
A Scientific Perspective on Big Data in Earth Observation
	1 Introduction
	2 Understanding the Challenges of Big Data from Space
	3 Deepen into Big Data Analysis
	4 Data Lifecycle and Developed Technologies
		4.1 Data Acquisition
		4.2 Data Organization, Management and Storage
		4.3 Data Processing and Analysis
		4.4 Provision of Value-Added Information
	5 Earth Observation Exploitation Platforms and Applications
		5.1 Overview of Platform Implementations
		5.2 Characteristics of the Platform Concept
	6 Public Awareness and Drawbacks
	7 Perspectives
		7.1 Market Perspectives
		7.2 European Strategic Interest
		7.3 Advancing the Scientific Development
	8 Conclusions
	References
Visualizing High-Dimensional Data Using t-Distributed Stochastic Neighbor Embedding Algorithm
	1 Introduction
	2 Literature Review
		2.1 Dimensionality Reduction Technique
		2.2 Background on Dimensionality Reduction and Feature Extraction
			2.2.1 Feature Extraction Techniques: Principal Component Analysis (PCA)
			2.2.2 Feature Extraction Techniques: t-Distributed Stochastic Neighbor Embedding (t-SNE)
	3 Algorithmic Details of t-SNE
		3.1 Algorithm
		3.2 Space and Time Complexity
		3.3 Effective Use of t-SNE
			3.3.1 Hyperparameter Changes the Whole Show
			3.3.2 Cluster Sizes in a t-SNE Plot
			3.3.3 Distances Between Clusters Might Not Mean Anything
			3.3.4 The Necessity of More Than One Plot for Topology
	4 t-SNE Versus PCA on MNIST Dataset
		4.1 Dimensional Reduction Using PCA on MNIST
		4.2 Dimensional Reduction on MNIST Using t-SNE
		4.3 t-SNE Usage
	5 Application of t-SNE
	6 t-SNE Versus PCA
	7 Common Fallacies of t-SNE
	8 Conclusion
	References
Active and Machine Learning for Earth Observation Image Analysis with Traditional and Innovative Approaches
	1 Introduction
	2 Data Set Description
	3 Ground Truth Data Generation
		3.1 Motivation
		3.2 Methodology
		3.3 Experimental Results
		3.4 Conclusions
	4 Normalized Compression Distance for Change Detection
		4.1 Motivation
		4.2 Methodology
		4.3 Experimental Results
			4.3.1 Danube Delta Scenario
			4.3.2 Belgica Bank Scenario
		4.4 Conclusions
	5 Extracting Features with Variational Auto-encoders
		5.1 Motivation
		5.2 Methodology
			5.2.1 Variational Inference and Auto-encoders
			5.2.2 Classifiers
			5.2.3 Overview of Our Methodology
		5.3 Experimental Results
			5.3.1 The Danube Delta Scenario
			5.3.2 The Belgica Bank Scenario
		5.4 Conclusion
	6 Coastline and Ice Detection Using Polarimetric SAR Images
		6.1 Motivation
		6.2 Methodology
		6.3 Experimental Results
			6.3.1 Danube Delta Scenario
			6.3.2 Belgica Bank Scenario
		6.4 Conclusions
	7 Outlook
	References
Applications in Financial Industry: Use-Case for Fraud Management
	1 Issues and Challenges for Application of Data Science Principles and Tools in Financial Industry
	2 A Data Science Use-Case for Financial Industry: Fraud Detection with Data Analytics and Machine Learning
		2.1 Fraud Management Solution Design Process with Supervised Learning
			2.1.1 Data Analytics in Fraud Management
			2.1.2 The Supervised Learning Process and Application in Fraud Management
		2.2 Fraud Management Solution Design Process with Unsupervised Learning and Anomaly Detection
	References
Stochastic Analysis for Short- and Long-Term Forecasting of Latin American Country Risk Indexes
	1 Introduction
	2 Problem Formulation
	3 Related Work
	4 Proposed Approach
		4.1 The Basic Problem
		4.2 NN-Based AR Model
		4.3 Monte Carlo Implementation Including Fractional Brownian Motion
		4.4 Algorithm Description
	5 Implementation with EMBI Time Series
		5.1 Monthly EMBI Forecast
		5.2 Using the Monthly EMBI Forecast
		5.3 Annual EMBI Forecast
		5.4 Using the Annual EMBI Forecast
		5.5 Discussion and Lessons Learned
	6 Conclusions
	References
Index