VDPC: Variational density peak clustering algorithm

The widely applied density peak clustering (DPC) algorithm makes an intuitive cluster formation assumption that centers are often surrounded by data points with lower local and far away from other higher density. However, this suffers one limitation it is problematic when identifying clusters because they might be easily merged into As a result, DPC may not able to identify variational To address issue, we propose (VDPC) algorithm, which designed systematically autonomously perform the task on datasets various types of distributions. Specifically, first novel method representatives among all construct initial based identified for further analysis clusters’ property. Furthermore, divide different levels according their unified framework combining advantages both DBSCAN. Thus, spreading across processed form final clusters. evaluate effectiveness proposed VDPC conduct extensive experiments using 20 including eight synthetic, six real-world, image datasets. experimental results show outperforms two classical algorithms (i.e., DBSCAN) four state-of-the-art extended algorithms.