Course Content Analysis: An Initiative Step toward Learning Object Recommendation Systems for MOOC Learners

With the accelerating development of open education, lowcost online learning resources, such as Massive Open Online Courses (MOOCs), are reaching a wide audience around the world. However, when faced with these appealing but overwhelming learning resources, learners are prone making rash learning decisions, which may be either excessive or insufficient to their learning capacities. To avoid the mismatch between learners and learning objects, we propose a supporting system that recommends a personalized path of learning objects for a given learner. In realizing this system, a domain knowledge structure is necessary to connect learners’ information and learning objects. As an initiative step, we employ the Labeled Latent Dirichlet Allocation method to predict how the content of a course is distributed over different categories in the domain. We conduct experiments by utilizing course syllabi as course content, and curriculum guidelines as domain knowledge. The predicting performance is improved when involving external texts related to the concerned domain knowledge unit.