Choosing the best resources: a recommender approach

May 21, 2018 10:40 AM – 11:30 AM

David Topps, University of Calgary
Tamsin Treasure-jones, University of Leeds
Sebastian Dennerlein, KNOW Center

Learners at all levels are deluged with open educational resources but receive little help in gauging which of these might be useful or applicable to their context. Traditional repositories rely on descriptors and metadata but learners seldom pore through these, relying on Google for recommendations.

A group of medical schools, in collaboration with a technical university in Austria, are seeking to integrate a Yelp-like recommender system that would help learners to better find resources that appeal to them, in much the same manner as we make restaurant recommendations to each other.

But much of the challenge is integrating educational activities and resources across the disparate systems of this international consortium. ETL (Extract, Transform, Load) approaches that have been typically recommended in such situations are way too expensive and inflexible to be feasible. The group is exploring the use of federated Learning Record Stores and xAPI to provide extractable data and linkable resources, without compromising the internal functionality of their existing systems.