Prediction; University Libraries; Book Loan;

The university library has become a longstanding institution within universities, playing a crucial role in supporting the teaching-learning process, research, and university extension. One of its most significant aspects is the loan of informational materials, especially books, to its users. Understanding the profile of users who make these loans and contextualizing their needs is essential for effective planning and optimized resource management within the library. In this context, this study aims to investigate the dynamics of book loans in the circulating collection of the Zila Mamede Central Library (BCZM), with the goal of anticipating user demands through the analysis of historical loan data. Initially, Exploratory Data Analysis was employed to understand relevant aspects of the interaction between students and the library, using loan data and information associated with students' academic lives. Clustering, performed through KNN and Hierarchical Clustering algorithms, allowed the identification of distinct student profiles, enriching the understanding of the specific needs of each group. Finally, to fully achieve the proposed objective, the Random Forest and SARIMA models were used to predict loans for the year 2019, using data grouped on a weekly and monthly basis. The results indicate that students who use loan services tend to have a higher academic completion rate and a lower incidence of withdrawals and cancellations. Regarding prediction models, both SARIMA and Random Forest proved promising in identifying trends in loans, highlighting their applicability to the Zila Mamede Central Library. This study not only contributes to a deeper understanding of the dynamics of book loans but also provides valuable insights for the library to proactively anticipate user needs, thereby improving the effectiveness of its operations.