<![CDATA[This study aims to develop a book recommendation system for a digital library using the collaborative filtering algorithm in order to address the problem of information overload and support users in finding relevant reading materials more efficiently. The research was conducted using a quantitative approach by collecting user–item interaction data, performing data preprocessing, and implementing both User-Based and Item-Based Collaborative Filtering models. The evaluation employed MAE, RMSE, Precision@K, and Recall@K to measure predictive accuracy and recommendation relevance. Experimental results show that Item-Based Collaborative Filtering outperforms the user-based model, achieving better accuracy and stability, particularly in highly sparse digital library datasets. The developed recommendation system demonstrates the ability to generate relevant book suggestions based on user behavior patterns, thus enhancing the overall user experience and information retrieval process. The findings highlight the potential of collaborative filtering as an effective method for improving digital library services and provide a foundation for future work involving hybrid approaches and strategies to mitigate cold-start challenges.]]>
Copyrights © 2026
