Learning Path Recommendation Using Lesson Sequence and Learning Object based on Course Graph

With the continuous popularity of online learning, it is difficult for learners to decide how to learn when they face a large number of learning resources, especially when they must balance the limited learning time available and multiple learning objectives under different learning scenarios. This paper presents a learning path recommendation using lesson sequence and learning object based on course graph. This paper tries to get a more personalized and suitable learning path from three aspects. First, this paper realizes the semantic association between knowledge points and resources by constructing a multi-relational course knowledge graph based on lesson layer with three-layered content hierarchy. Second, AprioriAll algorithm is applied to dig out the target knowledge point (start lesson) with the highest confidence level in the current knowledge point of the learning record from the learners' interactive data. Third, we provide flexible ways to continue learning by evaluating the learners' knowledge mastery, such as providing auxiliary learning paths to enhance the current knowledge. More importantly, the learner's interaction data and learning preferences are considered throughout the recommendation process, and some personalized parameters are allowed to be dynamically updated, which will make the recommendations more and more personalized and accurate with the increasing use.