Analysis of the Actviness Library Members Using K-Means Clustering

Abstract

University libraries play an important role in supporting academic activities, but the trend toward digital information access has made physical services less than optimal. At Bhayangkara University Jakarta Raya, data shows a difference between the frequency of visits and book borrowing, so user segmentation is needed. This study aims to group library members based on their level of activity using the K-Means Clustering algorithm, with variables of visit frequency and borrowing. The method used is quantitative with a data mining approach, utilizing secondary data from the library system for the period May–December 2024. The analysis process includes *data preparation*, modeling using *K-Means*, and evaluation using the *Davies-Bouldin Index (DBI)*. The results show that the optimal number of clusters is three, with a DBI value of 0.628, indicating that the cluster quality is quite good. The three clusters formed are: Active Borrowers (high visits and loans), Active Non-Borrowers (high visits, low loans), and Passive Members (low visits and loans). The uniqueness of this study lies in the simultaneous combination of two user behavior variables. This segmentation is useful as a basis for developing a performance portfolio and library service strategies that are more effective and tailored to user needs.