Enhancing E-Learning Materials Recommendations with Contextual Awareness and Collaborative Filtering

The number of learners using online learning materials has significantly increased in recent years. Due to the overload of information, many learners face difficulty finding relevant and useful educational materials that meet their needs. To solve this problem, recommendation systems (RS) are widely relied upon. These systems are used to provide suitable recommendations to users based on their needs. However, traditional recommendation techniques such as collaborative filtering (CF) and content-based filtering (CB) solely rely on user ratings and content attributes. Therefore, it is imperative to incorporate the learner's context when it comes to precise and personalized recommendations in e-learning. These conventional methods often neglect the learner's context when computing similarities and making recommendations, resulting in less accurate suggestions. This paper proposes a learning materials recommendation approach by integrating context awareness with collaborative filtering (CF) algorithms. Context awareness in the proposed approach entails incorporating context-specific data, like the learner's knowledge level and preferred learning style. Collaborative filtering computes predictions and generates targeted learner recommendations based on contextualized data. Evaluating the proposed recommendation approach (CF-CA) involved experiments comparing it with traditional collaborative filtering (CF) on a dataset of 349 learners. The two-recommendation technique's performance varied with neighbors (5 to 50). The proposed method demonstrated a 0.21 MAE improvement at an optimal size of 5. Furthermore, the proposed approach (CA-CF) consistently outperforms the CF in precision and recall across varying recommendation volumes. Additionally, by utilizing the learner's contextual data, the proposed method helps alleviate the cold-start issue.