Combining Content-based and Collaborative Filters in Learning Object Selection

Abstract

The explosive growth of online digital learning resources that commonly referred to as “learning objects” in elearning community, demands effective recommendation solutions. We explore the hybrid techniques, which are by combining the content-based with the conditions of the learning object attributes and collaborative filtering with the condition ofa learner similarity, to recommend the most suitable learningobject to learners. The preferred content-based algorithm andnearest neighbor-based algorithm has been developed, combiningwith ranking method to strengthen predictive results. A learning object raking example is discussed to demonstrate the method implementation by using setting experiment. This result shows that the combining method can reduce the error rate of learner dissatisfaction.