Gradient boosting algorithm for predicting student success

Abstract

The idea of using machine learning resolution techniques to predict student performance on an online learning platform such as Moodle has attracted considerable interest. Machine learning algorithms are capable of correctly interpreting the content and thus predicting the performance of our students. Algorithms namely gradient boosting machines (GBM) and eXtreme gradient boosting (XGBoost) are highly recommended by most researchers due to their high accuracy and smooth boosting time. This research was conducted to analyze the effectiveness of the XGBoost algorithm on Moodle platform to predict student performance by analyzing their online activities, practicing various types of online activities. The proposed algorithm was applied for the prediction of academic performance based on this data received from Moodle. The results demonstrate a strong correlation between many activities like the number of hours spent online and the achievement of academic goals, with a remarkable prediction rate of 0.949.