<![CDATA[This study aims to cluster social media user interaction patterns in order to obtain a more structured and informative mapping of online activities. User activity on social media platforms is increasing and diverse, ranging from the frequency of uploads, the intensity of comments, the pattern of information dissemination, to responses to certain content. This complexity creates problems in understanding the characteristics of user behavior as a whole, especially when the resulting data is very large and unstructured. To address this challenge, this study applies the K-Means Clustering algorithm, one of the data mining methods that is effective in clustering data based on similar characteristics. The dataset used comes from user activities that include the number of posts, the number of likes, the number of comments, and the level of daily interactions. K-Means is used to divide the data into several clusters that represent the types of user activities, such as active, semi-active, and passive users. The results show that the K-Means algorithm is able to produce a clear and measurable mapping of online activities, with evaluation values ??using SSE and silhouette scores indicating optimal cluster formation. From the grouping process, Cluster 1 was obtained as passive users consisting of U001, U002, U005, U007, and U009 with a final centroid value of C1 of (0.13; 0.13; 0.10; 0.18), Cluster 2 as active users consisting of U003, U004, U006, and U010 with a final centroid value of C2 of (0.53; 0.55; 0.59; 0.49; 0.62), and Cluster 3 as very active users consisting only of U008 with a final centroid value of C3 of (1.00; 1.00; 1.00; 1.00).]]>
Copyrights © 2026
