<![CDATA[This study proposes the implementation of the K-Means clustering algorithm to analyze and classify student graduation scores at SMK Negeri 3 Metro in order to support academic decision-making and improve information management. The dataset consists of academic records of students from grades X to XII, including theoretical scores, practical scores, attendance rates, and final achievement scores. Prior to clustering, the dataset is subjected to preprocessing procedures, including data cleaning, attribute transformation, and min-max normalization to ensure proportional scaling among variables. The clustering process is performed using RapidMiner and configured into three clusters representing categories of academic performance. The experimental results indicate that the K-Means algorithm effectively identifies structured patterns in the distribution of graduation scores, enabling the institution to map student achievement levels objectively. To operationalize the analytical results, the clustering model is integrated into a web/Android-based information system that facilitates real-time access to graduation information for teachers, administrators, and students. The implementation of the system contributes to improved efficiency in academic data management, faster information dissemination, and enhanced transparency in graduation scores. The findings demonstrate that the application of data mining techniques, particularly K-Means clustering, provides a reliable framework for academic analytics and decision support in educational information systems, with potential scalability for broader institutional adoption.]]>
