<![CDATA[Manual management of student financial transaction data remains a major challenge in many higher education institutions, including in the collaboration between Universitas Islam Negeri Imam Bonjol (UIN IB) Padang and Bank Nagari. Until now, no automated system has been developed to cluster student transaction data using the K-Medoids algorithm within higher education institutions in West Sumatra. This study aims to design a transaction clustering system that can identify student transaction patterns more efficiently. The K-Medoids algorithm is applied to transaction data that has been preprocessed through categorical transformation and normalization to address accuracy issues in distance-based analysis. The results show the formation of three main clusters: low (59 data points), medium (185 data points), and high (106 data points). This distribution reflects the variations in student transaction behavior and can be utilized by both the university and the bank to design more targeted service strategies, such as resource allocation and payment policy evaluation. This research provides an initial contribution to the application of K-Medoids-based data mining for optimizing transaction management in regional higher education institutions]]>
Copyrights © 2025
