Educational Data Science: Essentials, Approaches, and Tendencies: Proactive Education based on Empirical Big Data Evidence

Preface
Contents
Editor and Contributors
Part I: Logistic
Chapter 1: Engaging in Student-Centered Educational Data Science Through Learning Engineering
1.1 Introduction
1.1.1 Chapter Overview
1.1.1.1 State of the Art
1.2 Learning Engineering
1.2.1 Learning Engineering at Acrobatiq
1.3 Courseware in Context
1.3.1 Acrobatiq Courseware: Data-Driven
1.3.1.1 Learn by Doing Data
1.3.1.2 Data-Driven Adaptivity
1.3.1.3 Data for Instructor Feedback Loops
1.4 EDS to Investigate Learning by Doing
1.4.1 The Doer Effect
1.4.2 Summary
1.5 EDS for Evaluating AI-Generated Questions
1.5.1 AI-Generated Courseware
1.5.2 Courseware Adaptive Development
1.5.3 Courseware Implementation
1.5.4 Description of Analyses
1.5.5 Evaluation of AG Questions
1.5.5.1 Engagement
1.5.5.2 Difficulty
1.5.5.3 Persistence
1.5.5.4 Adaptive Analysis
1.5.6 Summary
1.6 EDS for Question Evaluation at Scale
1.6.1 The Content Improvement Service
1.6.2 Summary
1.7 Conclusion
References
Part II: Reviews
Chapter 2: A Review of Clustering Models in Educational Data Science Toward Fairness-Aware Learning
2.1 Introduction
2.2 Clustering Models for EDS Tasks
2.2.1 Methodology of the Survey Process
2.2.2 EDS Tasks Using Clustering Models
2.2.2.1 Analyzing Students´ Behavior, Interaction, Engagement, Motivation, and Emotion
2.2.2.2 Analyzing Students´ Performance and Grading
2.2.2.3 Predicting Students´ Performance
2.2.2.4 Supporting Learning, Providing Feedback and Recommendation
2.2.2.5 Supporting Collaboration Activities
2.2.2.6 Analyzing Physical and Mental Health
2.2.2.7 Miscellaneous Tasks
2.2.3 Clustering Models
2.2.3.1 -Means
2.2.3.2 -Medoids
2.2.3.3 Hierarchical Clustering
2.2.3.4 Fuzzy -Means Clustering
2.2.3.5 EM Clustering
2.2.3.6 SOM Clustering
2.2.3.7 Spectral Clustering
2.2.3.8 DBSCAN Clustering
2.3 Fair Clustering Models (for EDS)
2.3.1 Fairness Notions
2.3.1.1 Balance
2.3.1.2 Bounded Representation
2.3.1.3 Social Fairness
2.3.1.4 Individual Fairness
2.3.2 Fair Clustering Models
2.3.2.1 Fair Center-Based Clustering
2.3.2.2 Fair Hierarchical Clustering
2.3.2.3 Fair Spectral Clustering
2.4 Clustering Evaluation in EDS Toward Fairness-Aware Learning
2.4.1 Evaluation Measures
2.4.1.1 Clustering Quality Measures
2.4.1.2 Fairness Measures for Fair Clustering
2.4.2 Datasets
2.5 Beyond Fairness Requirements for Clustering in EDS
2.6 Conclusions and Outlook
Appendix
References
Chapter 3: Educational Data Science: An ``Umbrella Term´´ or an Emergent Domain?
3.1 Introduction
3.2 Review Settlement
3.2.1 Method
3.2.2 Resources
3.2.3 Taxonomy of Educational Data Science: A Proposal
3.3 Educational Data Science: A Glimpse
3.3.1 Context
3.3.1.1 Baseline
Big Data
Probability and Statistics
Data Analysis
Machine Learning
Data Mining
Knowledge Discovery in Databases
Analytics
Data Science
3.3.1.2 Related Field
Educational Big Data
Educational Data Mining
Learning Analytics
3.3.1.3 Incipient
Facade
Quotation
3.3.2 Profile
3.3.2.1 Introductory
Pioneer
Concepts
3.3.2.2 Scope
Umbrella
Concrete
3.3.2.3 Expression
Mention
Depth
3.3.2.4 Overview
Editorial
Review
3.3.3 Approach
3.3.3.1 Logistic
Frames
Frameworks
Conceptual
3.3.3.2 Applications
Instruction
Apprenticeship
Evaluation
Resources
3.4 Discussion
3.4.1 Analysis of Results
3.4.2 Reflection of Educational Data Science
3.4.3 Contribution of the Review
3.4.4 Responses to Research Questions
3.5 Conclusions
References
Part III: Applications
Chapter 4: Educational Data Science Approach for an End-to-End Quality Assurance Process for Building Creditworthy Online Cour...
4.1 Introduction
4.2 Credentialing in Distance Education
4.2.1 Background on Distance Education
4.2.2 Credentials in Distance Education
4.3 The Notion of Credits
4.3.1 Credit Hours
4.3.2 Quality Worthy of Credit
4.4 Massive Open Online Courses (MOOCs) for Credit?
4.4.1 Background on MOOCs
4.4.2 Assessing MOOC Quality
4.4.3 MOOC-Based Degrees and Credits
4.5 End-to-End Quality Assurance Process for Online Courses
4.5.1 Process Overview
4.5.2 Plan: Before Course Release-Catering to Different Learner Preferences
4.5.2.1 Sample 1: Measuring Content Type Distribution
4.5.2.2 Sample 2: Analyzing Content Type Sequence
4.5.3 Do: While the Course Is Running-Understanding Learner Engagement
4.5.3.1 Sample 1: Analyzing Activity for Different Video Types
4.5.3.2 Sample 2: Measuring Learner Sentiment on Discussion Forums
4.5.4 Check: Course Closure-Evaluating Goals Against Outcomes
4.5.4.1 Sample 1: Comparing Topics Discussed in the Course-Provided Materials and the Discussion Boards
4.5.4.2 Sample 2: Gauging the Acceptance of Automated Text-to-Speech Dubbing on Videos
4.5.5 Act: Course Revision-Aiming for Continuous Improvement
4.5.5.1 Sample 1: Identifying Metric Differences Across Course Revisions
4.5.5.2 Sample 2: Brainstorming from Course Feedback Survey
4.6 Other Quality Indicators
4.6.1 Accessibility: Readability and Listenability
4.6.2 Trustworthiness: Competency-Based Assessments That Are Hard to Cheat
4.6.3 Quality of Service: Timeliness of Feedback and Support
4.7 Conclusions
References
Chapter 5: Understanding the Effect of Cohesion in Academic Writing Clarity Using Education Data Science
5.1 Introduction
5.2 Related Work
5.2.1 Text Cohesion and Writing Clarity
5.2.2 Reverse-Cohesion Effect
5.2.3 The Role of Abstracts in Academic Writing Clarity
5.2.4 Measuring Text Cohesion Using Coh-Metrix
5.2.5 Hybrid Neural Fuzzy Inference System (HyFIS)
5.2.6 The Current Study
5.3 Methodological Framework
5.3.1 Data
5.3.2 Model Development
5.3.2.1 Data Preprocessing
5.3.2.2 Feature Extraction and Reduction
5.3.2.3 Clarity Score Classification Model
5.3.2.4 Performance Evaluation
5.4 Analysis and Interpretation of the Results
5.4.1 Descriptive Statistics and Feature Selection Results
5.4.2 Findings from the Classification Model
5.4.3 Findings from the Feature Weights and the Fuzzy Inference System
5.5 Discussion, Interpretations, and Findings
5.6 Limitations and Future Work
References
Chapter 6: Sequential Pattern Mining in Educational Data: The Application Context, Potential, Strengths, and Limitations
6.1 Introduction
6.1.1 Sequential Pattern Mining (SPM)
6.1.1.1 SPM Algorithms
6.1.1.2 An Example of Applying SPM to Synthetic Behavioral Sequences
6.2 What Modes of Educational Data and Research Purposes Is SPM Applicable to?
6.2.1 Data Modes
6.2.1.1 Event Logs
6.2.1.2 Discourse Data
6.2.1.3 Resource Accessing Traces
6.2.1.4 Other Data Sources
6.2.2 Research Purposes
6.2.2.1 Mining Learning Behaviors
6.2.2.2 Enriching Educational Theories
6.2.2.3 Evaluating the Efficacy of Interventions
6.2.2.4 Building Predictive Models
6.2.2.5 Developing Educational Recommender Systems
6.3 How Can SPM Contribute to Understanding the Temporal Aspects of Learning?
6.3.1 An Example: Combining SPM with Statistical Analyses to Understand Learning in an Open-Ended Learning Environment
6.3.1.1 Betty´s Brain
6.3.1.2 Discovering Similarities in Engagement and Learning
6.3.1.3 Discover Differences in Engagement and Learning
6.3.1.4 Discovering Changes in Engagement and Learning Over Time
6.3.1.5 Summary
6.3.2 Findings from Applications of SPM in Other Learning Environments
6.4 The Strengths, Limitations, and Future Directions of SPM in EDS
6.4.1 Strengths
6.4.1.1 Easy to Learn and Accessible
6.4.1.2 Revealing Meaningful Information That May Be Ignored Otherwise
6.4.1.3 A Powerful Tool for SRL Research
6.4.1.4 More Flexible than Other Sequential Analysis Methods
6.4.2 Limitations and Future Directions
6.4.2.1 No Available SPM Libraries for Computing Instance Values
6.4.2.2 Lack a Guideline for Preprocessing and Parameter Setting
6.4.2.3 Excessive Sequential Patterns
6.4.2.4 Interpreting Sequential Pattern Differences Is Challenging
6.4.2.5 Sequential Patterns Do Not Imply Causality
6.5 Conclusion
Appendix
R Code for the Synthetic Example
Python Code for the Synthetic Example
References
Chapter 7: Sync Ratio and Cluster Heat Map for Visualizing Student Engagement
7.1 Introduction
7.2 Review Baseline
7.2.1 Method
7.2.2 Background
7.2.3 Resources
7.2.4 Web Scraping for On-Site Learning Analytics
7.3 Sync Ratio of Opening Teaching Materials
7.3.1 Student Sync Ratio
7.3.2 Visualization Sync Ratio by Table and Graph
7.3.3 Each Material Sync Ratio and Learner´s Browsing Process
7.4 Student Engagement and Cluster Heat Map
7.4.1 Cluster Heat Maps and Outlier Detection
7.4.2 Taxonomy of Learning Patterns via Normal Distribution
7.4.3 Integration of Learning Patterns
7.4.4 Detection of Abnormal Values
7.5 Discussion
7.5.1 Effect of Real-Time Sync Ratio On-Site
7.5.2 Sync Ratio and Learner´s Reaction
7.5.3 Statistical Analysis Between Engagement Heat Map and Learning Patterns
7.5.4 Limitations of This Research
7.6 Conclusion
Appendix
References
Index