Hendrik Drachsler defended his PhD thesis on developing recommender systems to help learners find personalized paths through informal learning networks. His research included a theoretical background study, a psychology experiment, and a simulation of learning networks. The goal was to address the problem of learners feeling overwhelmed by the large amount of information in learning networks by creating a recommender system prototype that could suggest personalized learning activities and paths based on a learner's interests and previous activities. The conclusion was that recommender systems designed specifically for learning networks can help lifelong learners more efficiently pursue personalized learning goals.
5. Learning Networks
• explicitly address informal
learning
• allow learners to publish,
share, rate, tag and adjust
their own Learning Activities
their own Learning Activities
in a Learning Network
• contain open corpora that
emerge from the bottom
upwards
10. The PhD Project
j
Prototype:
Practical Recommender System
for Learning Networks
Study 3: Learning Networks
y g
Simulation
Study 2: Psychology Experiment
Study 2: Psychology Experiment
Theoretical Study 1: Theoretical Background
y g
2006 2007 2008 2009
hendrik.drachsler@ou.nl
Recommender Systems 2008, Lausanne
Page 10
11. The PhD Project
j
Prototype:
Practical Recommender System
for Learning Networks
2006
Study 3: Learning Networks
y g
Simulation
Study 2: Psychology Experiment
Study 2: Psychology Experiment
Theoretical Study 1: Theoretical Background
y g
2006 2007 2008 2009
hendrik.drachsler@ou.nl
Recommender Systems 2008, Lausanne
Page 11
12. The PhD Project
j
Prototype:
Practical Recommender System
for Learning Networks
2007
2008
2006
2009
Study 3: Learning Networks
y g
Simulation
Study 2: Psychology Experiment
Study 2: Psychology Experiment
Theoretical Study 1: Theoretical Background
y g
2006 2007 2008 2009
hendrik.drachsler@ou.nl
Recommender Systems 2008, Lausanne
Page 12
13. The PhD Project
j
Prototype:
Practical Recommender System
for Learning Networks
2008
2006
2009
Study 3: Learning Networks
y g
Simulation
Study 2: Psychology Experiment
Study 2: Psychology Experiment
Theoretical Study 1: Theoretical Background
y g
2006 2007 2008 2009
hendrik.drachsler@ou.nl
Recommender Systems 2008, Lausanne
Page 13
14. The PhD Project
j
Prototype:
Practical Recommender System
for Learning Networks
2006
2009
Study 3: Learning Networks
y g
Simulation
Study 2: Psychology Experiment
Study 2: Psychology Experiment
Theoretical Study 1: Theoretical Background
y g
2006 2007 2008 2009
hendrik.drachsler@ou.nl
Recommender Systems 2008, Lausanne
Page 14
15. Conclusions
Recommender Systems for learning
have to be designed differently to
recommender systems for e-
commerce.
Recommender Systems can support
lifelong learners to follow more
personalized learning paths. Further,
they positively influence the time they
need to reach their learning goals.