Identifying Critical Factors When Predicting Remedial Mathematics Completion Rates.

This article focuses on using machine learning (ML) techniques to predict student completion rates in remedial mathematics courses at a U.S. university serving many first-generation, low-income, and ethnic minority students. The study developed and compared seven ML models, finding the random forest ensemble model to be the most accurate with a 78.52% overall accuracy and 73.03% accuracy in predicting non-completion. Sensitivity analysis identified four key predictors of remedial mathematics completion: the remedial course where a student begins, credit completion rate, math placement score, and high school GPA, while demographic factors such as ethnicity and first-generation status were less influential. The findings suggest that ML models can support early identification of at-risk students to enable timely interventions, though further research is needed to improve model performance and incorporate additional predictive features.