双曲空间下一种基于节点坐标的快速网络社团检测方法
fast algorithm of detecting community structure based on coordinates of nodes in the hyperbolic space
本文分析了双曲空间中网络社团结构的表现特征并发现属于同一个社区的节点基本都被映射到了同一段角度范围内，高质量的社团结构在双曲圆盘上表现为扇区结构。进一步分析这种扇区形式的社团结构形成的机理，一种基于节点坐标信息的启发式快速优化模块度的社团检测方法(GOMA)被提出来。节点坐标信息的利用使得模块度能够快速收敛，极大的降低了算法时间复杂度。实验结果表明，该算法能检测到高质量的社团结构，并且该算法在应用于强聚集性的不均匀度分布的网络时效果更好。
his paper focus on the community structure of complex network embedded in hyperbolic space and verify that network communities manifest as sectors under the hyperbolic disk. Based on that, a fast algorithm of detecting community structure based on optimizing modularity making use of coordinates of nodes(GOMA) is proposed. Consideration of coordinates of nodes makes modularity of network convergent fast and reduce time complexity of the algorithm largely. The experiment results show that this algorithm can detect network community structure with high quality and is maximally efficient in networks with strongest heterogeneity and clustering.
王祖喜、熊威
计算技术、计算机技术
复杂网络社团结构双曲空间
complex networkcommunity structurehyperbolic space
王祖喜,熊威.双曲空间下一种基于节点坐标的快速网络社团检测方法[EB/OL].(2015-04-27)[2025-08-16].http://www.paper.edu.cn/releasepaper/content/201504-431.点此复制
评论