A Proposed Framework for Student’s Skills-Driven Personalization of Cloud-Based Course Content

Abstract

Engaging students’ personalized data in the aspects of education has been on focus by different researchers. This paper considers it vital for exploring the student’s progress, moreover, it could predict the student’s level which consequently leads to identifying the required student material to raise his current education level. Although the topic has been vital before the COVID-19 pandemic, however, the importance of the topic has increased exponentially ever since. The research supports the decision-makers in educational institutions as considering personalized data for the student’s educational tasks and activities proved the positive impact of raising the student level. The paper proposes a framework that considers the students’ personal data in predicting their learning skills as well as their educational level. The research included engaging five well-known clustering algorithms, one of the most successful classification algorithms, and a set of 10 features selection techniques. The research applied two main experiment phases, the first phase focused on predicting the students’ learning skills, and the second focused on predicting the students’ level. Two datasets are involved in the experiments and their sources are mentioned. The research revealed the success of the clustering and prediction tasks by applying the selected techniques to the datasets. The research concluded that the highest clustering algorithm accuracy is enhanced k-means (EKM) and the highest contributing features selection method is the evolutionary computation method.