Personalized Content Delivery in Content-Centric Online Learning

Personalization of online learning has become essential in the improvement of student satisfaction, motivation, and learning. With rapid growth of web-based systems like NPTEL students are introduced to thousands of courses, and this normally ends up in information overload. The proposed project is in the form of a content-based course recommender system that offers learners personalized content in an online learning platform that revolves around content. TF-IDF vectorization and cosine similarity are utilized in the system to provide recommendations through handling textual data in course names, descriptions, and instructors. Streamlit offers a user interface for the learner to engage through modules such as search, personal interests, departments wise grouping, and visualizations of datasets. Performance was measured based on metrics such as Precision@K, Recall@K, MAP, and NDCG, where TF-IDF was consistently superior to the baseline Count Vectorizer. This study indicates that content-based filtering has the potential to greatly enhance online learning by offering tailored recommendations synchronized with learning needs.