Hassan-2468/Predictive-Models-for-Student-Dropout-Risk-A-Review-and-Empirical-Evaluation-Using-Real-World-Data

This project evaluates machine learning models (XGBoost, Random Forest, and LR) to predict student dropout risk using the OULAD dataset. It demonstrates how behavioral data can serve as an early warning system to improve student retention, especially in low-resource environments like Sudan Project Overview : This project focuses on the critical issue of student dropout in higher education, specifically exploring predictive models as early warning systems. Key Highlights : The Problem : Addressing student attrition in challenging and low-resource environments (e.g., Sudan), where academic disruptions are common. The Data: Utilization of the OULAD (Open University Learning Analytics Dataset), incorporating demographic, academic, and behavioral interaction data. Models Used: Implementation and comparison of: Logistic Regression (Baseline) Random Forest (Ensemble) k-Nearest Neighbors (k-NN) XGBoost (Boosting) Results : The study confirms that ensemble and boosting techniques significantly outperform traditional models, with student interaction logs being the most influential predictors of success or dropout.