Clustering Library Loan Books Using K-Means Clustering

Optimal library collection management requires an understanding of book borrowing patterns to align availability with user needs. Without proper analysis, less popular books may remain in large quantities, while popular books may experience shortages. This study employs the K-Means Clustering method to group borrowed books at the Saintek UINSU Medan Library. The dataset consists of 290 loan records with attributes including book type, borrowing frequency, and the number of individuals borrowing each book. The data was converted into a numerical format and normalized using Min-Max Scaler. The Elbow Method was applied to determine the optimal number of clusters, which was found to be two. This study aims to classify books based on borrowing patterns to provide insights into library collection management. The clustering results can assist in decision-making regarding book procurement and distribution. Cluster C0 consists of popular books with high borrowing frequency and a large number of borrowers, while Cluster C1 includes books with lower borrowing rates. These findings offer a deeper understanding of borrowing trends, aiding libraries in developing acquisition strategies and organizing collections more effectively to meet user needs. These findings provide valuable insights for strategic decision-making in library collection development and maintenance, ensuring that popular books are adequately stocked while minimizing the accumulation of less-demanded titles.