Enhancing Collaborative Filtering-Based Course Recommendations by Exploiting Time-to-Event Information with Survival Analysis
Alireza Gharahighehi, Achilleas Ghinis, Michela Venturini, Frederik Cornillie, Celine Vens
TL;DR
This work addresses dropout-heavy MOOC environments by introducing a post-processing framework that injects time-to-event signals from survival analysis into collaborative filtering recommendations. By modeling time-to-dropout and time-to-completion with SA methods and re-ranking CF predictions, the approach yields improved top-$k$ recommendations on three public MOOC datasets, as evidenced by higher $C$-index and time-aware $NDCG$ measures. The study demonstrates that time-to-event information is informative for user preferences and engagement, with XGBoost-based survival models often delivering the strongest gains, especially when using combined dropout/completion signals. The findings suggest practical benefits for personalized learning in MOOCs and point to future work in ensembles and multi-task SA-CF frameworks to further enhance personalization and interpretability.
Abstract
Massive Open Online Courses (MOOCs) are emerging as a popular alternative to traditional education, offering learners the flexibility to access a wide range of courses from various disciplines, anytime and anywhere. Despite this accessibility, a significant number of enrollments in MOOCs result in dropouts. To enhance learner engagement, it is crucial to recommend courses that align with their preferences and needs. Course Recommender Systems (RSs) can play an important role in this by modeling learners' preferences based on their previous interactions within the MOOC platform. Time-to-dropout and time-to-completion in MOOCs, like other time-to-event prediction tasks, can be effectively modeled using survival analysis (SA) methods. In this study, we apply SA methods to improve collaborative filtering recommendation performance by considering time-to-event in the context of MOOCs. Our proposed approach demonstrates superior performance compared to collaborative filtering methods trained based on learners' interactions with MOOCs, as evidenced by two performance measures on three publicly available datasets. The findings underscore the potential of integrating SA methods with RSs to enhance personalization in MOOCs.