The increasing rate of student dropout in higher education institutions poses a significant challenge to educational success and resource allocation. This study focuses on early identification of at-risk students, aiming to predict early class dropout by analyzing key demographic, academic, and behavioral factors. Using a dataset from Kaggle, which includes various predictors of student retention, the research applies machine learning techniques such as Multi-Layer Perceptron (MLP), AdaBoost, Stacking Classifier, and Voting Classifier to identify students at risk of early dropout. The analysis reveals strong correlations between student performance, attendance, and re-enrollment behavior, emphasizing the importance of these factors for early intervention. By leveraging machine learning models, the study provides insights for timely and proactive strategies to improve student retention and reduce dropout rates.

Keywords: Early dropout prediction, student retention, machine learning, MLP, AdaBoost, Stacking Classifier, Voting Classifier, academic performance, attendance patterns, higher education, Kaggle dataset, intervention strategies.