[1]李赵兴,袁威龙*,李馨玲.融合引力公式的非重叠网络重叠社区检测算法[J].计算机技术与发展,2025,(06):1-9.[doi:10.20165/j.cnki.ISSN1673-629X.2025.0023]
　LI Zhao-xing,YUAN Wei-long*,LI Xin-ling.Gravitational Formula-based Overlapping Community Detection Algorithm in Non-overlapping Networks[J].,2025,(06):1-9.[doi:10.20165/j.cnki.ISSN1673-629X.2025.0023]
点击复制

融合引力公式的非重叠网络重叠社区检测算法()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2025年06期
页码:
1-9
栏目:
大数据与云计算
出版日期:
2025-06-10

文章信息/Info

Title:
Gravitational Formula-based Overlapping Community Detection Algorithm in Non-overlapping Networks
文章编号:
1673-629X(2025)06-0001-09
作者:
李赵兴; 袁威龙*; 李馨玲
榆林学院 信息工程学院,陕西 榆林 719000
Author(s):
LI Zhao-xing; YUAN Wei-long*; LI Xin-ling
School of Information Engineering,Yulin University,Yulin 719000,China
关键词:
重叠社区; 非重叠社区; 引力作用; 社区贡献; 复杂网络
Keywords:
overlapping communities; non-overlapping communities; gravitational force; community contributions; complex network
分类号:
TP311
DOI:
10.20165/j.cnki.ISSN1673-629X.2025.0023
摘要:
社区结构作为复杂网络的核心属性之一,扮演着关键角色。 通过社区检测技术深入剖析和解读复杂网络的架构与功能,对于揭示其内在机制和特性具有至关重要的意义。 相较于其他领域,非重叠社区发现研究成熟。 基于此,该文提出了一种基于融合引力公式的非重叠网络重叠社区检测算法。 该算法借鉴了物理学中的引力概念,将网络拓扑信息融入万有引力公式,以此评估节点间的相互作用力,进而推导出节点与社区的作用力,以确立各社区的潜在成员节点集合。 依据节点对社区的贡献度,筛选出最终的重叠节点。 实验结果表明,该算法在真实网络 Lesmis 上的 EQ 值比 CPM 算法高出约 175. 75% ,在部分人工合成网络上的 ENMI 接近于 1,显示出该算法在真实网络和人工网络上均具有良好的性能,为网络重叠社区检测领域提供了一种新的研究思路。
Abstract:
Community structure plays a key role as one of the core attributes of complex networks. It is of vital significance to deeply analyze and interpret the architecture and function of complex networks through community detection techniques to reveal their intrinsic mechanisms and properties. Compared with other fields,the research on non-overlapping community discovery is mature,based on which an overlapping community detection algorithm for non - overlapping networks based on the fusion gravity formula is proposed. The algorithm draws on the concept of gravitational force in physics and incorporates network topology information into the gravitational force formula as a means of assessing the interaction force between nodes,and then deriving the force between nodes and communities in order to establish the set of potential member nodes of each community. The final overlapping nodes are filtered based on their contribution to the community. The experimental results show that the EQ value of the proposed algorithm is about 175. 75% higher than that of the CPM algorithm on the real network Lesmis,and the ENMI is close to 1 on some synthetic networks. It is showed that the proposed algorithm has excellent performance on both the real network and the artificial network,and provides a new research idea in the field of detecting the overlapping communities of the network.

相似文献/References:

[1]王庚,宋传超,盛玉晓,等.基于标签传播的社区挖掘算法研究综述[J].计算机技术与发展,2013,(12):69.
　WANG Geng,SONG Chuan-chao,SHENG Yu-xiao,et al.Research Summary on Communities Mining Algorithm Based on Label Propagation[J].,2013,(06):69.
[2]方木云,刘洪彬,谢恩文. Hadoop下基于边聚类的重叠社区发现算法研究[J].计算机技术与发展,2015,25(03):58.
　FANG Mu-yun,LIU Hong-bin,XIE En-wen. Research on Overlapping Communities Detecting Algorithm Using Hadoop Based on Edge Clustering[J].,2015,25(06):58.
[3]韩红旗,徐紫燕,李琳娜,等.重叠社区发现算法评价指标综述[J].计算机技术与发展,2024,34(05):1.[doi:10.20165/j.cnki.ISSN1673-629X.2024.0033]
　HAN Hong-qi,XU Zi-yan,LI Lin-na,et al.Overview on Evaluation Indicators for Overlapping Community Discovery Algorithms[J].,2024,34(06):1.[doi:10.20165/j.cnki.ISSN1673-629X.2024.0033]

更新日期/Last Update: 2025-06-10