ADAPTIVE DENSITY PEAK CLUSTERING FOR DETERMINGING CLUSTER CENTER

Density peak clustering algorithm is a new algorithm to achieve fast clustering by finding density peak. It has the advantages of simple implementation, less parameters required, processing non convex data and good clustering effect. However, it also relies on the truncation distance dc and can not automatically identify the cluster center. In this paper, a density peak clustering algorithm is proposed to adaptively determine the clustering center. First, the new local density is determined by the new neighborhood relationship, and then the clustering center is determined by the original algorithm to complete the clustering. Finally, the accuracy is verified by experiments.