Abstract | ||
---|---|---|
Recently, density peaks clustering (DPC) has received much attention. The key step in DPC is to find the nearest denser point (NDP) for each point. If a point and its NDP are close, they are in the same cluster; otherwise, they are in different clusters. However, the density gap between a point and its NDP is not considered in DPC. Once the point and its NDP are close, they are assigned to the same cluster, even if the density gap between them is large, which could result in poor clustering results. In this study, we propose a clustering algorithm based on finding prominent peaks (prominent peak clustering, PPC). Its main concept is to divide points into multiple potential clusters and then merge clusters whose density peaks are not prominent to obtain accurate clustering results. The prominence of cluster density peaks is measured by the density gap. Experimental results on low- and high-dimensional datasets demonstrate that PPC is competitive with other clustering techniques. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1016/j.engappai.2019.07.015 | Engineering Applications of Artificial Intelligence |
Keywords | Field | DocType |
Clustering,Density estimation,Density peaks,Density gap | Cluster (physics),Mathematical optimization,Computer science,Algorithm,Cluster analysis,Merge (version control) | Journal |
Volume | ISSN | Citations |
85 | 0952-1976 | 0 |
PageRank | References | Authors |
0.34 | 0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Li Ni | 1 | 8 | 4.18 |
Wenjian Luo | 2 | 356 | 40.95 |
Wenjie Zhu | 3 | 3 | 1.40 |
Wenjie Liu | 4 | 0 | 0.34 |