<![CDATA[As centers of literacy and learning, libraries face challenges in understanding book lending patterns to meet the needs of diverse users. The main problem faced is the lack of data-based analysis in optimizing library services and collections. This research aims to classify book borrowing patterns based on profession using the Naive Bayes algorithm, utilizing data from the Sidoarjo Library Service in 2023. The data consists of 4476 transactions with attributes such as profession, book category, and level of reading interest. This research was conducted in several phases, namely data collection preprocessing, processing using Gaussian and Multinomial Naive Bayes algorithms, and model evaluation. By testing on various data ratios (90:10, 80:20, 75:25, and 50:50), the results show that Gaussian Naive Bayes provides the highest accuracy of 97% in the random dataset scenario. The main findings show that students, university students and housewives dominate the high reading interest category, while doctors and researchers have lower reading interest. The unique value of this research is in its application of. data-based analysis to support library management. The research results provide strategic insight for developing more responsive data-based services, optimizing collections according to professional needs, and increasing the effectiveness of literacy programs. This research is anticipated to serve as the initial phase in utilizing data mining technology to overcome modern challenges in library management.]]>
