Recommendation strategies for e-learning: preliminary effects of a personal recommender system for lifelong learners

Hendrik Drachsler, Hans Hummel, Bert Van den Berg, Jannes Eshuis, Adriana Berlanga, Rob Nadolski, Wim Waterink, Nanda Boers, Rob Koper

    Research output: Contribution to conferencePaperAcademic

    18 Downloads (Pure)

    Abstract

    This article presents research on personal recommender systems for lifelong learning. The personal recommender systems supports lifelong learners in Learning Networks. A first version was evaluated in an experiment during an Introduction Psychology course of the Psychology Department at the Open University of the Nederlands. The learning activities of the psychology course and the personal recommender system were integrated into a Moodle environment, which operates as an emulated Learning Network. Therefore, no curriculum structure was applied and the students were allowed to study the learning activities in any order they wanted. The implemented personal recommender system combines a top-down, ontology-based recommendation technique with a bottom-up, stereotype filtering technique. Both techniques were combined in a recommendation strategy that decided which of the techniques were most suitable for the current situation a learner was in. This article presents preliminary results of the experiment and discusses the advantages and disadvantages of the used recommendation strategy. It further argues for the benefit of recommendation strategies for a personal recommender system in e-learning in general.
    Original languageEnglish
    Publication statusPublished - 15 Jul 2007

    Keywords

    • navigation
    • Personal Recommender System
    • collaborative filtering
    • ontologies
    • recommendation strategies
    • learning networks

    Fingerprint

    Dive into the research topics of 'Recommendation strategies for e-learning: preliminary effects of a personal recommender system for lifelong learners'. Together they form a unique fingerprint.

    Cite this