Utilizing Artificial Neural Network to Analysis of Academic Performance and Retention in Civil Engineering Education

The increasing demand for evidence-based approaches in education has led to the exploration of artificial intelligence techniques for predicting student performance and retention. This study utilizes Artificial Neural Networks (ANN) to forecast Civil Engineering students' academic outcomes, addressing the growing concern over student attrition. Given the rising dropout rates, early identification of at-risk students is crucial to implementing timely intervention strategies. This research constructs an ANN-based predictive model using key academic metrics, including quiz and exam scores, to analyze their correlation with student success. The model, developed using MATLAB, demonstrated strong predictive capability with an optimal configuration of four hidden layers, achieving <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$R$</tex>-values of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$0.82,0.78$</tex>, and 0.75 for training, validation, and testing datasets. The model accurately classified 68 % of student outcomes, providing a reliable tool for early risk detection. Findings underscore the significance of ANN in educational analytics, offering a data-driven approach to student retention. By integrating this model into academic support systems, institutions can enhance student success through personalized interventions such as tutoring and counseling. Future improvements include incorporating additional academic and behavioral indicators to refine prediction accuracy further.