Online courses recommendation system using content-based filtering / Nik Muhammad Irfan Shamsul Zuhairi

A content-based filtering approach serves as the primary basis for developing this project's main recommendation functionality for online courses. Users currently face challenges in finding proper online courses from numerous choices because the new system design establishes an improvement. The content-based filtering system personalizes course ideas based on the attributes of the courses and the user's preferences, improving the learning experience by making relevant and efficient recommendations. In this project, the objectives are to study the implementation of content-based filtering, develop a prototype system, and evaluate its performance. This approach enables users to find courses matching their interests with ease, hence enhancing their learning experience. The growing number of Massive Open Online Courses (MOOCs) available online makes it difficult for students to choose courses that align with their interests and learning objectives. To solve this, an online course recommendation system that employs Content-Based Filtering (CBF) is presented. The system personalizes recommendations based on course attributes and user preferences, improving the learning experience. The methodology involves collecting data from Kaggle, preprocessing it, then using TF-IDF and Cosine Similarity to calculate course similarity scores. A prototype system was developed which allowed users to provide their course option and receive personalized recommendations. The system's performance was evaluated using Mean Absolute Error (MAE) and Root Mean Square Error. The results show that the system has an MAE of 0.25 and an RMSE of 0.35, showing its ability to provide recommendations. Future work will include implementing hybrid filtering techniques, real-time user feedback, and increasing the dataset with additional diverse course sources to increase recommendation performance.