Predicting Student Success in MOOCs using Machine Learning: A Comparative Study of Classification Models

The proliferation of Massive Open Online Courses (MOOCs) has completely changed the educational landscape by providing students with access to high quality education content worldwide. However, issues like high dropout rates and uneven student interest levels still exist despite MOOCs’ growing popularity. This paper explores the potential of artificial intelligence (AI) in evaluating and enhancing the learning process within the framework of MOOCs. We present an empirical examination of MOOC data with an emphasis on dropout trends, course completion rates, and student engagement. We forecast student achievement using machine learning models, such as Random Forest classifiers, by taking into account factors like quiz scores, platform usage, and engagement levels. Our findings show that Gradient Boosting has achieved better results in terms of accuracy and AUC. Additionally, we pinpoint important classes with high dropout rates and examine the underlying causes, providing information on how AI might be used to improve student retention. The study highlights the benefits of AI-based prediction models for individualized instruction, giving teachers the means to proactively assist and intervene on behalf of pupils who are at risk. These results highlight the important role AI can play in improving learning outcomes in the MOOC ecosystem, which will increase student retention and success.