Hybrid Filtering Recommendation in E-Learning Environment

Abstract

Personalized recommendation in an e-learning system can actively introduce useful learning resources for learners. It is a “push” mechanism in contrast to the “pull” way like Web searching. At the same time it is also a very efficient way especially when users can not describe their needs exactly. This paper put forward an approach to recommend right learning resources for users with different learning needs by hybrid filtering method. Learning resources are organized by learning topics through text analysis. Users with similar learning interests are found out to form different common interest groups by user behavior tracing and recording. Then, two-level user profiles are built based on common interest group detection and text analysis. At last, learning resources are introduced to users according to user profiles by collaborative filtering and content-based filtering respectively. A time factor is also introduced into the building of user profiles, which makes user profiles adapt to user’s interest shifting.