The Density-graded Clustering Algorithm

In recent years, the widely used Density Peak Clustering (DPC) algorithm has given a clustering strategy in which the center of clusters is usually surrounded by data objects with locally lower densities, but it also results in some otherwise low-density clusters not being clearly distinguished due to such a definition. In this paper, we propose a density hierarchical clustering strategy for the density characteristics of the dataset itself, and divide the dataset according to the density level to form a subset, and different subsets respond to different strategies, which can effectively solve the above problems. Finally, the method is applied to six datasets, all of which have achieved good results.