Analyzing dropout of students and an explainable prediction of academic performance utilizing artificial intelligence techniques

Abstract

Modern higher education institutions (HEIs) face significant challenges in identifying, students who are at risk of low academic performance, at an early stage, while maintaining educational quality, and improving graduation rates. Predicting student success and dropout is crucial for institutional decision-making, as it helps formulate effective strategies, allocate resources efficiently, and improve student support. This study explores machine learning (ML) models for predicting student success, focusing on predicting first-semester CGPA (Cumulative Grade Point Average) and identifying at-risk students. It aims to compare various classifiers and regression models, identify the most effective techniques, and provide explainable insights into the decision-making process using Explainable AI (XAI). The results suggest that Logistic Regression outperforms other models in predicting at-risk students with high precision and recall, offering a reliable tool for early interventions.