Research on Topic Mining of MOOC Course Reviews Based on the BERTopic Model

This study uses the BERTopic model to conduct an in-depth analysis of MOOC course reviews, aiming to identify and optimize key curricular features that impact the learning experience. It has been found that despite the rapid development of the MOOC platform in China, there is a disconnect between curricular design and learner’s needs, which affects rates of course completion. The BERTopic model, with its superior semantic understanding, captures and interprets deeper themes in the data more accurately than traditional topic models. In this study, the model was applied to MOOC reviews and successfully identified 50 topics, showing higher topic consistency compared to existing models. Morever, the BERTopic model reveals possible course issues hidden in themes with a low number of comments, suffering data support and practical guidance for MOOC curricular improvement and educational innovation.