A Cluster Based Recommendation System for MOOCs

Massive Open Online Courses (MOOCs), which give students access to a variety of courses from renowned universities and knowledgeable professors, have completely changed how education is delivered. As there are so many alternatives, it may become difficult for learners to find and choose courses that suit their needs both academically and personally. Uncertainty and bewilderment may result from this overwhelming scene. Finding relevant courses, evaluating the quality of the content, and reviewing how well the courses match with long-term objectives are only some of the challenges. This lack of self-assurance may lead to less-than-ideal experiences, untapped potential, and discontent, which may have an effect on both individual students and educational institutions. By offering learners individualized recommendations based on their preferences and learning profiles, course recommendation systems have become an important tool to address this issue. This research aims to develop a course recommendation system using cluster analysis to group similar courses based on attributes like course ID, title, and description. The system uses a two-step approach: collecting an extensive dataset of MOOCs, preprocessing it, and applying the K-means Clustering algorithm to identify similar groups. This personalized recommendation system can help learners discover relevant MOOCs and meet their individual preferences and goals. The research focused on recommending similar courses using a dataset of 14,952 records. Clustering algorithms and the cosine similarity metric were used to group similar courses together. The model was compared to expert recommendations for each course.