A Hybrid Tag-Based Recommendation Mechanism to Support Prior Knowledge Construction

Prior knowledge in concept acquisition among students is an important issue. Traditional studies on prior knowledge generation during learning have focused on extracting sentences from reading materials that are manually generated by website administrators and educators. From the reports of previous studies, tag-based recommendation and assessment has been recognized as being an effective approach that can assist learners in finding out the clues and concepts of articles, such that it enables learners to be familiar with the learning content. However, sparse noisy data influences the quality of recommendations, especially for tag choices that are not mapped to the features of both the users and content. To cope with these problems, we adopt a hybrid recommendation method to consider tag preference, tag relevance and social networking. The experimental results show that the approach benefits from the additional information embedded in social knowledge, and can be an effective and efficient mechanism for enhancing the quality of prior knowledge recommendation.