Research On Resource Recommendation Using Tags On E-Learning Platforms

With the rapid development of the Internet technology and the increasing emphasis on online learning, E-Learning(Electronic Learning) platforms have been widely used. At the same time, with the growing number of learning resources, it is more and more difficult for learners to find the resources which can meet their needs in the traditional resource organization.Personalized recommendation technology which is able to actively recommend interesting information to users is currently one of the most effective ways to solve the information overload problem. Social tags can be used to organize and manage resources in a user-centered way. Users can tag on the resources which they are interested in with freedom so that users can organize resources in a personalized way and show their latent interests preference. Therefore, the introduction of social tags and personalized recommendation technology on E-Learning platform, not only can enrich the resource organization with social tags, but also can recommend the personalized learning resources according to the tagging information.In order to solve the resource management problems on current E-Learning platforms, social tags and personalized recommendation technology used as knowledge background, tag-based personalized recommendation algorithm viewed as research object, to improve the experience of resource search taken as research purpose, the author’s major contribution are outlined as follows:1. The existing tag-based resource recommendation algorithm models have been studied and summarized. Then the most widely used extended collaborative filtering model is studied as a research focus based on this summarization.2. In order to reduce the noise of the tag-based collaborative filtering algorithm caused by the semantic ambiguity and redundancy of social tags, firstly the popular tags are selected by the wisdom of crowds to profile users and resources. Then a collaborative filtering algorithm is designed based on the popular tags.3. The traditional tag-based collaborative filtering algorithm and the proposed popular-tag-based collaborative filtering algorithm have been compared on the Movie Lens 10 M dataset. The experiment shows that the proposed algorithm has reduced the noise of tags and improved the precision of the traditional tag-based collaborative filtering algorithm.4. “Embedded School” has been studied as an example of E-Learning platforms. To solve the organization problem of “Embedded School”, a social tag system has been designed and completed. The completed tag system which can be used to manage and organize resources, not only enrich the current resource organization but also provide the original data for personalized recommendation.5. A tag-based recommendation framework has been designed based on the completed tag system and the proposed popular-tag-based collaborative filtering algorithm.