Enhancing Library Resource Recommendations Using Collaborative Filtering Algorithms

Abstract

 The exponential growth of digital content in libraries poses significant challenges for users searching for relevant resources. This paper presents a collaborative filtering (CF)-based recommendation system designed to provide personalized recommendations for books, journals, and articles in digital library environments. The system leverages user-based and item-based CF models, enhanced by a hybrid approach that combines CF with content-based filtering (CBF) to address challenges like data sparsity and cold-start issues. Using a dataset of library interactions, the hybrid model demonstrates superior performance with an accuracy of 88% precision and a diverse catalog coverage of 92%, significantly improving user satisfaction and engagement. Experimental results show the system’s scalability and efficacy, underscoring the potential of hybrid recommendation models to enhance user experience in academic libraries.