A Semantic Role Mining and Learning Performance Prediction Method in MOOCs

Massive Open Online Courses (MOOCs) have gained tremendous popularity in the last few years. Discussion forums are a common element in MOOCs. Previously, Numerous studies have been undertaken on the potential role that discussion forums play in education. However, the existing works don’t effectively mine and analyze the rich text information of the forums associated with learner performances. In this work, we propose a hybrid method to mine learner roles based on MOOC discussion forums and to jointly evaluate the quality of learning with other learning activities. We pay more attention to extracting semantic features of posts and comments in forums, which help to promote the performance prediction. We evaluate the performance of our method based on the Coursera platform. Experiments show that our approach can improve the performance compared to existing works on these tasks.