Cone Cluster Labeling for Support Vector Clustering

Clustering forms natural groupings of data points that maximize intra-cluster similarity and minimize intercluster similarity. Support Vector Clustering (SVC) is a clustering algorithm that can handle arbitrary cluster shapes. One of the major SVC challenges is a cluster labeling performance bottleneck. We propose a novel cluster labeling algorithm that relies on approximate coverings both in feature space and data space. Comparison with existing cluster labeling approaches suggests that our strategy improves efficiency without sacrificing clustering quality.