The analysis of vessel sailing patterns is crucial for
enhancing maritime safety and navigation efficiency. This study
compares the performance of K-Means and DBSCAN clustering
techniques in analyzing vessel movements using space-based
Automatic Identification System (AIS) data from the LAPAN-
A2 and LAPAN-A3 satellites, focusing on eastern Indonesian
waters. Vessel stability is assessed through the standard
deviation of Speed Over Ground (SOG) and Course Over
Ground (COG). The AIS data is normalized using Min-Max
scaling and analyzed with both K-Means and DBSCAN
algorithms. Performance is measured using the Silhouette Score
and the Calinski-Harabasz Index to evaluate cluster quality.
Results show that K-Means achieves a Silhouette Score of 0.552
and a Calinski-Harabasz Index of 814.846, identifying stable
movement patterns in well-defined clusters. Conversely,
DBSCAN achieves a higher Silhouette Score of 0.731, better
detecting anomalies and noise, though with a lower Calinski-
Harabasz Index of 361.665. This comparative analysis
underscores the complementary strengths of each method,
offering valuable insights for maritime authorities in improving
navigational safety and operational efficiency through advanced
data analytics.
