Penerapan Data Mining Dalam Peminjaman Buku Baca Untuk Umum Berdasarkan Tren Peminjaman Di Perpustakaan Sumatera Barat

Book borrowing trend analysis is an important aspect of supporting library collection management and improving library services. The West Sumatra Library manages a large amount of borrowing data; however, these data have not been optimally utilized to identify borrowing patterns and classify books based on their popularity.. Therefore, this study aims to develop a web-based data mining application by implementing the K-Means clustering algorithm to analyze book borrowing trends at the West Sumatra Library. The study utilizes book borrowing data consisting of three attributes, namely total borrowings, weekly borrowing frequency, and average borrowing duration These attributes are processed using the K-Means algorithm to classify books into three clusters: Highly Popular, Moderately Popular, and Less Popular. The system was developed using the PHP programming language and MySQL database to facilitate data management, clustering calculations, and the presentation of analysis results automatically. The clustering process includes data preparation, centroid initialization, Euclidean distance calculation, cluster assignment, centroid updating, and iteration until convergence is achieved. The results of this study indicate that the developed system is capable of grouping books according to their borrowing characteristics accurately and efficiently. The generated clustering information can support librarians in identifying highly demanded books, optimizing collection management, and assisting decision-making related to book procurement and collection development. The results of this study show that the implemented K-Means algorithm successfully groups books into three categories: Highly Demanded, Moderately Demanded, and Less Demanded based on their borrowing characteristics. The system provides an accurate, efficient, and consistent analysis while reducing the time required compared with manual processing. These results can support library management in evaluating collections, improving services, and planning future collection strategies.