Density Peak Clustering Algorithm Considering Topological Features

• DOI: 10.3390/electronics9030459
• Publication Date: 2020-03-09T09:37:34Z
• Authors: Shuyi Lu; Yuanjie Zheng; Rong Luo; Weikuan Jia; Jian Lian; Chengjiang Li
• Keywords: Data stream clustering; Single-linkage clustering

The clustering algorithm plays an important role in data mining and image processing. breakthrough of precision method directly affects the direction progress following research. At present, types algorithms are mainly divided into hierarchical, density-based, grid-based model-based ones. This paper studies Clustering by Fast Search Find Density Peaks (CFSFDP) algorithm, which is a new based on density. has characteristics no iterative process, few parameters high precision. However, we found that did not consider original topological data. We also similar to social network nodes mentioned DeepWalk, satisfied power-law distribution. In this study, tried graph algorithm. Based previous studies, propose adds basis CFSFDP Our experimental results show with features significantly improves effect proves addition effective feasible.