Information entropy peaks clustering using dynamic reverse nearest neighbor sequence and 3D decision graph

doi:10.1016/j.eswa.2025.128197

科研成果详情

题名	Information entropy peaks clustering using dynamic reverse nearest neighbor sequence and 3D decision graph
作者	Lu，Jianyun 1,2; Chen，Zhong 3; Shao，Junming 1,4,5; Wu，Chunling 2
发表日期	2025-09-01
发表期刊	Expert Systems with Applications
ISSN/eISSN	0957-4174
卷号	288
摘要	Density Peaks Clustering has been widely studied due to its ability to detect clusters with arbitrary shapes and sizes. However, inappropriate density measures can result in the detection of multiple density peaks within a single cluster, leading to a degeneration of the availability of density peaks clustering. In this paper, we propose a new concept, information entropy peaks, which is robust to variable densities and noise and can accurately identify cluster representatives. To achieve this, we introduce a dynamic reverse nearest neighbor sequence to compute information entropy for each data point. We then develop a 3D decision graph that utilizes information entropy, Euclidean distance, and cosine distance to identify potential cluster representatives. The cosine distance is calculated based on the dynamic reverse nearest neighbor sequence and can measure distance in the direction of space. Finally, a nearest neighbor propagation algorithm is applied to form clusters, initialized with cluster representatives. Our method is effective in identifying cluster representatives and is robust to variable densities and noise. Comprehensive experiments on both synthetic and real-world datasets have demonstrated that our method produces better clustering results compared to many state-of-the-art clustering algorithms.
关键词	3D Decision graph Clustering Consine distance Dynamic reverse nearest neighbor sequence Information entropy peaks
DOI	10.1016/j.eswa.2025.128197
URL	查看来源
语种	英语English
Scopus入藏号	2-s2.0-105005872991
引用统计
文献类型	期刊论文
条目标识符	https://repository.uic.edu.cn/handle/39GCC9TT/13434
专题	个人在本单位外知识产出
通讯作者	Shao，Junming
作者单位	1.School of Computer Science and Engineering,University of Electronic Science and Technology of China,Chengdu,Sichuan,China 2.School of Artificial Intelligence and Big Data,Chongqing Polytechnic University of Electronic Technology,Chongqing,China 3.School of Computer Science,Peking University,Beijing,China 4.Shenzhen Institute for Advanced Study,University of Electronic Science and Technology of China,China 5.Yangtze Delta Region Institute (Quzhou),University of Electronic Science and Technology of China,China
推荐引用方式 GB/T 7714	Lu，Jianyun,Chen，Zhong,Shao，Junminget al. Information entropy peaks clustering using dynamic reverse nearest neighbor sequence and 3D decision graph[J]. Expert Systems with Applications, 2025, 288.
APA	Lu，Jianyun, Chen，Zhong, Shao，Junming, & Wu，Chunling. (2025). Information entropy peaks clustering using dynamic reverse nearest neighbor sequence and 3D decision graph. Expert Systems with Applications, 288.
MLA	Lu，Jianyun,et al."Information entropy peaks clustering using dynamic reverse nearest neighbor sequence and 3D decision graph". Expert Systems with Applications 288(2025).