A comparative study of machine learning algorithms for virtual learning environment performance prediction

Abstract

<p>Virtual learning environment is becoming an increasingly popular study<br />option for students from diverse cultural and socioeconomic backgrounds<br />around the world. Although this learning environment is quite adaptable,<br />improving student performance is difficult due to the online-only learning<br />method. Therefore, it is essential to investigate students' participation and<br />performance in virtual learning in order to improve their performance. Using<br />a publicly available Open University learning analytics dataset, this study<br />examines a variety of machine learning-based prediction algorithms to<br />determine the best method for predicting students' academic success, hence<br />providing additional alternatives for enhancing their academic achievement.<br />Support vector machine, random forest, Nave Bayes, logical regression, and<br />decision trees are employed for the purpose of prediction using machine<br />learning methods. It is noticed that the random forest and logistic regression<br />approach predict student performance with the highest average accuracy<br />values compared to the alternatives. In a number of instances, the support<br />vector machine has been seen to outperform the other methods.</p>