A Data Driven Approach in Foreasting Student Outcomes with Feed Forward Neural Networks

Predicting student learning outcomes is essential for identifying students who may need additional support and for improving educational strategies. This research introduces a predictive model that combines a Feedforward Neural Network (FNN) and Random Forest to forecast student performance based on structured data, including academic, behavioral, and demographic factors. The model evaluates key variables such as attendance, assignment grades, exam scores, and individual student profiles to predict whether students will meet their academic goals. Its efficient, lightweight design allows for fast data processing, while the integration of Random Forest enhances robustness and decision reliability, making the system adaptable to different educational environments. Additionally, the combined approach provides clear and interpretable results, helping educators pinpoint the factors that affect student performance and take targeted actions to improve outcomes. This methodology highlights how deep learning and ensemble learning can drive data-driven decisions, supporting better educational practices and student success.