AI-Enhanced Personalized Learning in Higher Education: Integrating Machine Learning With Moodle

Abstract

Learning Management Systems (LMSs) have been rapidly adopted in higher education. The emergence of the COVID-19 pandemic not only disrupted normal, traditional teaching and learning but also revolutionized pedagogical approaches and learning methodologies. However, most LMS platforms, such as Moodle, utilize static content delivery methods that are not sufficiently catered to the diverse learning needs and abilities of students. The study, motivated by the results from our previous article based on a systematic literature review, aimed to design and evaluate an AI-enabled adaptive learning framework to enhance Moodle’s ability to provide personalized learning experiences. This paper used a quantitative experimental research design and a data set that contained demographic, behavioral, engagement, and academic performance data from 213,000 students. The proposed framework involved the use of Extreme Gradient Boosting (XGBoost) models for learner performance prediction and at-risk student identification, along with a hybrid recommendation system for personalized content delivery. The results showed a remarkable predictive power, with the at-risk detection model achieving 97.7% accuracy and the learner performance prediction model achieving 97.3% accuracy. The feature importance analysis suggested that the top predictors of academic outcomes included learner engagement, peer interaction, attendance, time spent on the platform, and academic skills. Moreover, the hybrid recommendation system was able to generate personalized learning resources based on learners’ performance, deficiencies, and learning preferences. The research results indicate that the combination of AI and machine learning techniques in Moodle can greatly enhance adaptive learning, learner engagement, and early intervention strategies, resulting in improved educational outcomes in higher education settings.