Building group recommendations in e-learning systems

  • Authors:
  • Danuta Zakrzewska

  • Affiliations:
  • Institute of Information Technology, Technical University of Lodz, Lodz, Poland

  • Venue:
  • Transactions on Computational Collective Intelligence VII
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Building groups of students of similar features enables to suggest learning materials according to their member needs. The paper presents an agent-based recommender system, which, for each new learner, suggests a student group of similar profiles and consequently indicates suitable learning resources. It is assumed that learners can be characterized by cognitive styles, usability preferences or historical behavior, represented by nominal values. Building recommendations by using a Naïve Bayes algorithm is considered. The performance of the technique is validated on the basis of data of learners, who are described by cognitive traits such as dominant learning style dimensions or by usability preferences. Tests are done for real data of different groups of similar students as well as of individual learners.