Research and Development of An Innovative System to Fore case Secondary School Students Academic Accomplishments

Abstract

Academic performance prediction is a critical aspect of educational data mining (EDM), enabling educators, policymakers, and institutions to implement data-driven strategies for student success. This research presents an innovative predictive system that leverages Machine Learning (ML) techniques to forecast students' grades and percentage based on a diverse range of academic, socio-economic, psychological, and infrastructural factors collected through a structured survey. The dataset encompasses academic scores, attendance, study habits, parental support, mental health status, school facilities, technology accessibility, and government education initiatives. A Random Forest model is employed due to its ability to handle complex, heterogeneous data while providing insights into the most influential factors affecting student performance. Additionally, feature selection techniques are applied to extract key determinants, enabling personalized interventions. The proposed system not only enhances the accuracy of academic predictions but also provides actionable insights to bridge learning gaps, improve student engagement, and optimize resource allocation. This research significantly contributes to educational analytics, offering a scalable and interpretable model that aids in fostering academic excellence through intelligent, data-driven decision-making.