Early Prediction of Student Dropout in E-Learning Environments Using AI-Based Predictive Analytics: A Combined Theoretical and Practical Study

Student dropouts in virtual learning environments e-learning are a major headache for educational institutions as it affects the retention rate and the allocation of resources. The research combines theoretical knowledge gained from the AI-driven predictive models literature with an actual analysis of a student dataset from higher education studying the retention. The study work has identified through exploratory data analysis (EDA) and correlation studies the main factors that have a significant influence on dropouts such as financial problems, academic performance in the first few semesters, and demographic variables. Various machine learning models, such as Random Forest, XGBoost, and CatBoost, are put to work in prediction, thus revealing the possibility of an intervention. The research results show the presence of strong correlations between the performance metrics of each semester and also suggest that AI analytics can be quite accurate in forecasting the underprivileged students leading to the allocation of support resources in advance

https://doi.org/10.65723/RMSP2635