<![CDATA[Analyzing students’ academic achievement patterns is important to support data-driven academic decision-making. This study aims to identify academic achievement patterns of Informatics students at Universitas Hamzanwadi using an unsupervised learning approach based on the K-Means algorithm. The main contribution of this research lies in conducting scenario-based experiments using several combinations of academic attributes and applying multi-metric evaluation to determine the most optimal clustering configuration for seventh-semester student data. The dataset includes the Semester Grade Point Average (IPS), Cumulative Grade Point Average (IPK), and the total accumulated credit units (SKS). After data cleaning and preprocessing, 382 student records were obtained for analysis. Data normalization was performed using Min–Max Scaling. The optimal number of clusters was determined using the Elbow method, resulting in three clusters. Cluster quality evaluation produced a Silhouette Score of 0.71, a Davies–Bouldin Index of 0.54, and a Calinski–Harabasz Index of 808.64, indicating good clustering quality. The results revealed three groups of students with high, medium, and low academic achievement characteristics. These findings confirm that the K-Means algorithm is effective in clustering student academic achievement patterns and can be used as a basis for academic evaluation and student development strategies. However, this study is limited to seventh-semester data and can be extended to other datasets or across study programs]]>
