Evaluating Boosting Algorithms for Academic Performance Prediction in E-Learning Environments

Abstract

As educational institutions and platforms get more advanced, predicting academic success is becoming a lot easier. The goal of this research is to find the best boosting algorithm for academic prediction. To do that, we need to go through the data and analyze it. Three popular boosting algorithms, AdaBoost, HistGradientBoosting, and CatBoost were used for comparison analysis and also want a full performance measure assessment. The big winner is CatBoost. In terms of accuracy, precision, recall, F1 score, and AUC score it outperformed the competition by a long shot. Something truly impressive that CatBoost did was identifying students who may be at risk and decreasing positive predictions so they can focus on areas that need help without being overwhelmed. This study shows how important it is to choose the right algorithm in classifying tasks. Early prediction in e-learning is an important factor in the rapidly changing field of education as technology improves especially since it has potential to improve how students learn and support them with resources.