SIRLiB – Book Recommendation System in Libraries: A book recommendation system to promote reading

Abstract

A book recommendation system aims to facilitate readers' access to works that match their interests and personal preferences. The primary objective of such a system is to create an efficient mechanism that suggests books based on users' previous behavior and preferences, thereby increasing satisfaction and engagement in reading. The main goal of this work was to develop such a system to recommend books of interest to users. This system was named SIRLiB . The theoretical contributions of this work include the study of recommendation systems, especially those based on clustering and content-based filtering. Additionally, data mining techniques such as the K-Means algorithm and the TF-IDF algorithm were explored to identify patterns in user data and book attributes. First, K-Means was applied to categorize the inserted books, and after this phase, TF-IDF was used to recommend these books. The methodology adopted involved collecting reading preference data from some users, followed by the application of recommendation algorithms. The main results achieved demonstrated an improvement in the accuracy of the recommendations, with the system successfully suggesting relevant books to most users.