Sustaining Student Academic Supervision Using Information Gain and Random Forest Classifier

Sustaining academic supervision faces challenges, such as time constraints, large intake of students, inefficient use of technology, etc. These challenges can negatively impact the quality of academic supervision, leading to issues such as student’ dissatisfaction, decreased learning outcomes, and higher student dropout rates. Thus, a proper tool to address the challenges is required. This paper discusses the use of a combination of the Information Gain feature selection method with the Random Forest classifier to provide more accurate features or factors that can improve learning achievement. The experiment results on a dataset collected from the learning management system of the university show promising accuracy results, with the best accuracy achieved being 89.90%, compared to 80.50% achieved by the J48 classifier with manual feature selection.