Research on the universal evaluation method of MOOC courses based on machine learning

The prevailing MOOC evaluation metrics are deficient in their inability to assess the quality of courses that are pervasive in nature. In this paper, we propose a machine learning model-based method for evaluating the pervasiveness of MOOC courses. The proposed method first utilizes learning data on the MOOC platform to obtain the pattern between effort and learning effect. Then, it uses this pattern to quantify the pervasive quality of MOOC courses in terms of difficulty and required knowledge reserve. The method does not necessitate the administration of additional questionnaires to obtain feedback evaluations, and it is capable of discerning valid relationships between students' learning behaviors and learning outcomes. The metrics obtained can be utilized to inform instructors' curricular reforms and students' course selection. The proposed method is validated using learning data from the edX MOOC platform, which is freely accessible to the public.