Advancing Education: Hybrid Recommendation Systems for Best-Fit Student Domain Matching
Sarra Aouadi, Toufik Marir, Mohammed Lamine Kherfi · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2024
Universities around the world are concerned with the student dropout phenomenon, which is particularly prevalent in the early years.Research indicates that the main reason for early dropout is the wrong choice of academic study domain.In this work, we have tried to provide decision-making support to the new students to help them choose the path that best suits their abilities and skills.From a conceptual perspective, we propose a hybrid recommendation system that integrates machine learning algorithms and collaborative filtering techniques to address real-world educational big data.From a practical standpoint, this system utilizes the machine learning model to identify the academic domain in which a student is most likely to succeed.Subsequently, collaborative filtering is applied to utilize the top 20% of similar students to estimate potential success rates within the predicted domain.Our approach introduces several significant innovations compared to existing methods, demonstrating improved prediction accuracy and offering the potential to positively impact academic success rates