A Privacy-Preserving Framework Enhancing University Student Engagement Using Machine Learning and Gamification

Student disengagement and high attrition continue to be a major problem in higher education, particularly in intensive disciplines like computer science. These challenges are often compounded by problems such as academic stress, mental health issues and the particular adjustment difficulties of international students. Traditional teaching methods often lack the flexibility and responsiveness required to meet the diverse and evolving needs of learners. In response, this paper presents a comprehensive conceptual framework that combines machine learning techniques with a structured gamification design to improve student engagement. The proposed model consists of three interrelated layers: The first involves systematic data collection and feature engineering of virtual learning environments, capturing behavioural indicators such as interaction with content and activity on the platform. The second layer applies unsupervised machine learning algorithms to create dynamic engagement profiles that enable continuous monitoring and identification of student engagement patterns. The third layer uses the Web-Agon framework to design and adapt gamification strategies based on these profiles and offer personalised interventions tailored to individual engagement. The framework also incorporates privacy-preserving technologies, including federated learning and differential privacy, to ensure ethical handling of sensitive data. This approach aims to create an adaptive and student-centred learning environment that promotes sustained engagement and academic success.