Xiao Hu "Overview of the Space of Learning Analytics and Educational Data Mining"
1. Overview of the Space of
Learning Analytics and
Educational Data Mining
George Siemens (University Of Texas, Arlington And
SoLAR) & Ryan Baker (Columbia University)
Presented by Xiao Hu @ University of Hong Kong
2. Scholarly Communication
Venues Provided by SoLAR
• LAK, JLA, LASI, LAMP
• The mailing-lists
• LA Google group:
https://groups.google.com/forum/#!forum
/learninganalytics
• EDM-Discuss
http://www.educationaldatamining.org/IE
DMS/mailinglists
3. Types of EDM Methods
• Prediction
• Structure Discovery
• Relationship Mining
• Distillation of Data for Human Judgment
• Discovery with Models
• Bring SoL and analytic geeks together
4. What's Going on in Learning
Analytics
• Research methods are becoming finely
turned
• Growing impact on practice
• Increased Computer Science presence
5. What's Going on in EDM
• Lots of classification and regression
• Increased emphasis on latent
knowledge estimation and knowledge
structure discovery
•figure out what students doing
• Reduced emphasis on relationship
mining (association rules)
6. What's Going on in EDM
(Con’d)
• Constructs studied are broadening
• Meta-cognition
• Affect
• Engagement
• Motivation
• Long-term participation (cross-years)
• Increased number of studies with models
analysis, particularly on more generalized
models
7. What's going on in EDM
(Con’d)
• Basic research
• Automated intervention
• Getting into "reporting“
•understand the meaning of the model,
tell people what we found
• Increased participation of industry
8. Going Forward…
• New tools, techniques and people (disciplines)
• Data: openness, ethics and scope
• Target of analytics activity
• Connections to related fields and practitioners
(collaborations, LAMP)
• Challenges: creating strategic links to related
communities
9. Coming Up
• A MOOC on EdX "Data, Analytics and
Learning", Starting Oct. 20th, 2014
• LAK2015: in NY (deadline: Dec. 24th)
• ICEDM2015 in Madrid (deadline: Dec. 24th)
10. Q1: Friction between Computer
Scientists and Learning Analytics ?
• LAK vs EDM papers?
•EDM don't use huge datasets
• Around what is knowledge, fairness,
equality
• Will see more and more studies using CS
methods, but the attributes of learning
experience are also important
11. Q2: Collaboration with Start-
ups?
• Certainly open to building connections to
start-ups, as there is so much activity going
on
• But such connections are not yet explicit
• Companies are scattered, connections are
much appreciated!
12. Q3: Research Ethics : Consent
form, Privacy etc. ?
• Do students have to give consent?
• Not clear, no legal architecture in this area
• Privacy: does it apply to MOOCs?
• As a community, we need to educate the IRB
about the nature of our research;
• Improve the ways of anonymizing data;
• Educate ourselves about how to manage data
(for long-term access and re-use)