Time-Based Recommendations for Lecture Materials
Christoph Hermann · 2010
Abstract: In this paper we describe the implementation of a system for improving the utilization of lecture materials through recommendations. We present results of an evaluation of the usage of lecture materials as well as an evaluation of download-statistics. These findings led us to develop a novel approach for a time based recommender system for lecture materials. Our approach to making recommendations is particularly useful for Learning Management Systems or Lecture Archives. It differs significantly from previous work due to the fact, that we are employing the timestamps from usage data to create a similarity measure for items. The algorithm of this recommender is presented. We evaluate our algorithm in comparison with other recommendation algorithms like a Loglikelihood recommender for boolean preferences. We show that our algorithm has a better performance then a Loglikelihood recommender.