A Hybrid AI Approach for Predicting Academic Performance in RBE Students

Willy Gonzales¹Zindel Cordero¹Carlos D. Abanto-Ramírez¹Elías A. Torres Armas²Hasnain Iftikhar³Javier Linkolk López-Gonzales^1*

• ¹Universidad Peruana Union, Lima District, Peru
• ²Universidad Nacional Toribio Rodriguez de Mendoza de Amazonas, Chachapoyas, Peru
• ³Quaid-i-Azam University, Islamabad, Pakistan

Machine learning has advanced significantly in recent years and is being used in higher 3 education to perform various types of data analysis. While the literature demonstrates the 4 application of machine learning algorithms to predict performance in university education, no 5 such applications are found in EBR, let alone in private institutions of a denominational nature, 6 which presents an opportunity to study prediction in these institutions. To address this gap, this 7 research aims to propose a predictive approach as a decision-support tool for regular basic 8 education, using machine learning techniques. Among the techniques utilized, three machine 9 learning models (Logistic Regression, Support Vector Machine, and Random Forest), along 10 with deep learning models (AlexNet, Gated Recurrent Unit, and Bidirectional Gated Recurrent 11 Unit), were analyzed, as well as ensemble models. Nonetheless, the Ensemble model, which 12 combines deep learning and machine learning techniques, is preferred due to its superior 13 accuracy, precision, and sensitivity performance metrics.