Development Methodologies for Big Data Analytics Systems: Plan-driven, Agile, Hybrid, Lightweight Approaches (Transactions on Computational Science and Computational Intelligence)

✍ Scribed by Manuel Mora (editor), Fen Wang (editor), Jorge Marx Gomez (editor), Hector Duran-Limon (editor)

• Publisher: Springer
• Year: 2023
• Tongue: English
• Leaves: 289
• Category: Library

✦ Synopsis

This book presents research in big data analytics (BDA) for business of all sizes. The authors analyze problems presented in the application of BDA in some businesses through the study of development methodologies based on the three approaches – 1) plan-driven, 2) agile and 3) hybrid lightweight. The authors first describe BDA systems and how they emerged with the convergence of Statistics, Computer Science, and Business Intelligent Analytics with the practical aim to provide concepts, models, methods and tools required for exploiting the wide variety, volume, and velocity of available business internal and external data - i.e. Big Data – and provide decision-making value to decision-makers. The book presents high-quality conceptual and empirical research-oriented chapters on plan-driven, agile, and hybrid lightweight development methodologies and relevant supporting topics for BDA systems suitable to be used for large-, medium-, and small-sized business organizations.

✦ Table of Contents

Editorial Preface
References
Acknowledgments
Contents
About the Editors
Contributors
Open Source IT for Delivering Big Data Analytics Systems as Services: A Selective Review
1 Introduction
2 Background
2.1 Foundations of Big Data Analytics Systems
2.2 Models for Implementing and Delivering IT Services
2.3 The NIST Big Data Reference Architecture (NBDRA)
3 Selective Review of Open Source IT for Implementing and Delivering BDASaaS
4 Discussion of Contributions
5 Conclusions
Appendix
References
The Role of Machine Learning in Big Data Analytics: Current Practices and Challenges
1 Introduction
2 Machine Learning Techniques
2.1 Support Vector Machines
2.2 Decision Trees
2.3 Clustering Algorithms
2.4 Artificial Neural Networks
3 Open-Source Platforms for Big Data Analytics
3.1 MapReduce
3.2 Apache Hadoop
3.3 Apache Spark
3.4 Other Open-Source Platforms and Tools
4 Domain Areas of Big Data Analytics
4.1 Healthcare
4.2 Weather Forecasting
4.3 Social Networks and the Internet
5 Conclusions
References
The Data Value Chain Ontology
1 Introduction
2 Problem Identification and Motivation
3 Definition of Objective and Solution
4 Methodology
4.1 Results of Initial Literature Review and Workshop
4.2 Results of the Evaluation Model
4.3 Results of the Extended Literature Review
5 Ontology Derived from the Results of the Literature Review
5.1 Delimitation
5.2 Ontology Structure and Taxonomy
5.3 Data Product
5.4 Infrastructure Management
5.5 User Interface
5.6 Evaluation Model
6 Visualization
7 Discussion and Outlook
References
Requirements for Machine Learning Methodology Software Tooling
1 Introduction
2 Method: From Stakeholders to Requirements Capture
3 Machine Learning Process Models
3.1 KDD
3.2 SEMMA
3.3 CRISP-DM
3.4 CRISP-ML(Q)
3.5 Data-to-Value (D2V)
4 Requirements for Software Support Tools
4.1 Overall Vision
4.2 User Stories and Requirement Templates
4.3 From Requirements Toward OO Classes/ER Entities
5 Related Work
5.1 Work on Tooling for Machine Learning Process Methodology
5.2 Requirements Capture
5.3 Workbenches for Constructing Machine Learning Pipelines
5.4 Work on Support Tooling for Business Workflows
5.5 Work on Support Tooling for Machine Learning Development, Deployment, and Operations
6 Discussion
7 Summary, Conclusions, and Future Work
References
A Selective Conceptual Review of CRISP-DM and DDSL Development Methodologies for Big Data Analytics Systems
1 Introduction
2 Research Method
3 Background
3.1 Foundations of Big Data Analytics Systems (BDAS)
3.2 The ISO/IEC 29110 Standard – Basic Profile – as a Lightweight Development Process Template
4 Selective Comparative Analysis of BDAS Development Methodologies
4.1 Description of the Rigor-Oriented CRISP-DM Methodology
4.2 Description of the Lightweight DDSL Methodology
4.3 Comparison of the Rigor-Oriented CRISP-DM and the Lightweight DDSL Methodologies
5 Discussion of Contributions and Conclusions
5.1 Discussion of Contributions
5.2 Conclusions
Appendix (Figs. 10, 11, 12 and 13)
References
A Selective Comparative Review of CRISP-DM and TDSP Development Methodologies for Big Data Analytics Systems
1 Introduction
2 Research Methodology
3 Background
3.1 Foundations of Big Data Analytics Systems (BDAS)
3.2 The Scrum-XP Workflow: An Agile Framework of Practices
4 Selective Comparative Analysis
4.1 Description of the CRISP-DM Methodology
4.2 Description of the TDSP Methodology
4.3 Comparison of the CRISP-DM and TDSP Methodologies
5 Discussion of Contributions and Conclusions
5.1 Discussion of Contributions
5.2 Conclusions
References
BDAS-EPM: An Integrated Evolution Process Model for Big Data Analytics Systems
1 Introduction
2 Background Overview
3 Selective Review Research Method
4 Results and Synthesis
4.1 The Main Concepts and Evolution of BDA
4.2 The Most Relevant BDA Frameworks
4.3 Applications of BDA
4.4 BDA Challenges and Trends
4.5 Illustration and Discussions on the BDAS-EPM
5 Conclusion
References
Big Data Adoption Factors and Development Methodologies: A Multiple Case Study Analysis
1 Introduction
2 Background
3 Literature Review
3.1 Big Data
3.2 Previously Studied Big Data Adoption Factors
3.3 Development Methodologies
4 Methodology
5 Procedure
6 Participants
7 Data Analysis
8 Validity
9 Findings
9.1 Big Data Adoption Findings
The Challenge of BD Value: New Insights
The Challenge of Security (Old, New, and Unique)
Challenge of Managing Large Datasets
More Data Means More Privacy Concerns
Cost of Big Data
The Burden of Regulations
IT Expertise in Big Data
Big Data Adoption Findings Summary
9.2 Big Data Development Methodology Findings
Medium Development Team Size Is Common
Agile Development Method Is the Popular Option
Agile Development Has Drawbacks Too
Big Data Development Methodology Findings Summary
10 Study Limitations and Future Research Opportunities
11 Conclusion
References
Detection of Breast Cancer in Mammography Using Pretrained Convolutional Neural Networks with  Fine-Tuning
1 Introduction
2 Previous Works
3 Material and Methods
3.1 CNN Architectures
VGG19
ResNet-50 and ResNet152
EfficientNetB7
3.2 Datasets
3.3 Experiment Environment
4 Methodology
4.1 Stage 1
4.2 Stage 2
4.3 Stage 3
4.4 Stage 4
4.5 Preprocessing
5 Results and Evaluations
5.1 Metrics
5.2 Tables
5.3 Comparison with Previous Works
6 Conclusions and Future Work
References
Challenges and Opportunities of Intercompany Big Data Analytics in Supply Chains
1 Introduction
2 State of the Art of Supply Chain Data Exchange
3 Challenges for Big Data Integration in Supply Chain
4 Benefits of Intercompany Big Data Analytics
4.1 Management and Planning
4.2 Logistics
4.3 Production
4.4 Discussion
5 Technical Concepts to Connect Data Stores Securely
5.1 Data Spaces
5.2 GAIA-X
5.3 Federated Learning
6 Discussion and Further Work
References
From Big Data to Big Insights: A Synthesis of Real-World Applications of Big Data Analytics
1 Introduction
1.1 Characteristics of Big Data
2 Application of Big Data Analytics in the Healthcare Industry
3 Application of Big Data Analytics in the Retail Industry
4 Application of Big Data Analytics in the Telecommunication Industry
5 Implications for Research and Practice
5.1 Future Research Directions
6 Conclusion
References
Index