DDCF: Enhancing Educational Resource Recommendation in E-Learning Platforms Using Collaborating Filter Approach

An e-learning-based recommendation system suggests learning resources, courses, or materials to users based on their preferences, behavior, and historical interactions within an e-learning platform. This system aims to enhance the learning experience by providing personalized and relevant content to individual users. However, lacking feedback can make improving and fine-tuning the challenging recommendation algorithms difficult. Additionally, inaccurate or insufficient metadata can lead to less accurate recommendations. To improve recommendation systems in e-learning, we propose a novel deep-learning technique. Data is collected from the Udemy dataset and then processed—cleaning the data and normalizing it to perform well in the recommendation system. Relevant features are extracted to classify and recommend courses to the user. The Hierarchical Multi-scale Long Short-Term Memory (HMLSTM) technique is used to classify the course level into beginning, intermediate, and advanced categories. The user’s ratings are then identified to recommend courses using the Improved Golden Jackal Optimization (IGJO) Algorithm. Finally, courses are recommended to the user based on user ratings using a collaborative filtering technique. The performance of the proposed method is evaluated by precision, accuracy, recall, and F1-score metrics and compared to state-of-the-art techniques.