Abstract: Classical clustering algorithms such as k-means face limitations in handling clusters with heterogeneous shapes, densities, and sizes, while exhibiting sensitivity to initial centroid ...