In recent years we have seen a dramatic change in the education world towards wide-adoption of online learning technologies (e-learning). The huge amount of fine-grained data being collected by these environments can be used to study human learning, improve the design of interactive learning environments, and develop data-driven analytics solutions that will assist students, teachers, content developers, and administrators.
These fall into the emerging fields of learning analytics and educational data-mining (EDM). The course will cover the main research topics within this domain, and will also touch upon some of the technological issues that are relevant for conducting such research.
At the end of the course, students should be familiar with the key research directions and methodologies in the domain, understand the potential of EDM tools to their scientific discipline (e.g., Physics Education), and hopefully, how they can be applied to their specific research.
Below is a list of various topics that the course will touch upon (not necessarily in this order). The topics are organized by the domain and the methodology.
Domain (goals, processes, big questions, conceptual frameworks):
- Learning Processes: measuring learning, Identifying effective learning paths, in-video behavior
- Cognitive Model Discovery, and using it to optimize the design of interactive learning environments
- Contexts: Open-ended learning environments, MOOCs, Intelligent Tutoring Systems
- Dashboards and Visualization: Communicate learning analytics to teachers and other stakeholders
- Ineffective and anti-learning behaviors and strategies – ‘Gaming the system’, cheating, boredom, and more.
- Adaptive Learning and Content Recommendation
- Student at-risk and drop-out prediction
- Collaborative and social learning
- Privacy, Security, and ethics
Methods and Technology:
- Data Sources and types: Clickstream data, text (e.g. forum discussions), Assessment, multi-modal.
- Supervised and Unsupervised machine learning
- Natural Language Processing, Text Mining, and application for Assessment, Sentiment Analysis, and open questions
- Bayesian Knowledge Tracing (BKT), Hidden Markov Models (HMM)
- Cognitive Modeling, Q-matrix construction, Exploratory Factor Analysis (EFA), Item Response Theory (IRT)
- Gaming The system and Cheating
- Recommender Systems
- Social Network Analysis
- Simulating data with HMM and IRT
- Video analysis
- Sequence Mining, Temporal Analysis
- Sensitivity Analysis, Anomaly Detection
The course will be in the form of a seminar, with two presentations during the semester, and a final assignment. For the final assignment, students could choose between a theoretical/hands-on assignment.