PREDICTING STUDENT ATTRITION VIA ENSEMBLE MACHINE LEARNING: IMPLICATIONS FOR ENROLLMENT MANAGEMENT

Abstract

Losing grade point average is the most stressful issue under which teens and even their future career. It hurts so many people in general. Research in how a student quits their education at some point.  More complex data models could help identify exactly why an employee wants to leave, possibly because they are already happy where they're at. Other people might eat more efficiently if everyone started at lunch. Researchers apply machine learning to make educated guesses about students' futures based on their demographics and behaviors. Compared to other classification methods, which ones outperform the rest? By comparing which method is more efficient we can use that information to plan for future admissions and kind of perfect it. Because XGBoost is better than a other algorithm XGBoost has 89 percent accuracy and an 0.93 area of a curve Monitoring students closely will help schools understand that students have negative problems that affect their passage through life and correct them so the student will see a positive change. Policy makers by giving advice and ideas can improve our systems and make them better programs to do.

Keywords: student attrition, ensemble machine learning, predictive analytics, enrollment management, higher education retention, Random Forest, XGBoost