K-Means Algorithm with Aftershock Validation for Earthquake Clustering in North Maluku

North Maluku has several tectonic earthquake sources, which cause high seismic activity, so understanding earthquake clustering patterns is very important for seismic risk and hazard assessment. Validate the clustering by comparing the resulting clusters with the distribution of aftershocks concentrated around the main earthquake source. The cluster will be better if the cluster found matches the aftershock pattern. This research focused on the application of K-Means algorithm, which is an unsupervised machine learning clustering technique, to analyze the spatial patterns of earthquakes in North Maluku based on 8549 earthquake data in the time frame of 2019-2024. The clustering result were compared using 538 aftershock sequence data from three major earthquakes that occurred in the region. K-Means algorithm processes data to be closely related to the centroid when grouping data into k clusters, the value of k is determined using elbow method and cluster quality metrics index optimization. K-Means clustering successfully identified 4 earthquake clusters, with quality metrics of Silhouette Coefficient 0.28, Calinski-Harabasz Index 4458.84, Davies-Bouldin Index 1.16. Validation with aftershock data shows that 9 8 % to 1 0 0% of aftershocks are in one of the seismic zone clusters.