基于重叠社区发现的网络数据可视化优化方法研究与实现  

Research and Implementation of Network Data Visualization Optimization Method Based on Overlapping Community Discovery

作　　者：解蓝莹 周莲英[1] 谢超 XIE Lanying;ZHOU Lianying;XIE Chao(School of Computer Science and Communication Engineering,Jiangsu University,Zhenjiang 212013)

机构地区：[1]江苏大学计算机科学与通信工程学院,镇江212013

出　　处：《计算机与数字工程》2024年第2期477-481,577,共6页Computer & Digital Engineering

基　　金：慢病本体知识库开发项目(编号:8421170004);市智慧妇幼信息平台本体知识库系统项目(编号:20180477)资助。

摘　　要：伴随数据的迅猛增长,数据间关系变得错综复杂,给网络数据可视化带来了挑战。通过社区发现,凸显网络中的局部聚类特性可以提高可视化效果,而重叠社区的发现更贴近现实中的网络结构。具有简单高效执行速度快的Louvain算法是目前最常用的社区发现算法之一,但重叠社区的发现是其不足之处。为此,论文以Louvain算法为基础,结合基于谱映射的模糊C-means聚类算法改进社区发现算法,改进的算法利用谱映射将数据节点映射到欧几里得空间,以隶属度计算数据节点属于某个聚类的程度,由此可以允许同一数据属于多个不同的类,从而实现重叠社区结构的发现,最后基于所提出改进算法,使用主流布局算法中的FR模型对网络数据进行可视化。以模块度值作为评估指标,实验结果表明,论文提出的方法能够发现重叠社区,可以清晰地展示网络中的社区结构,在经典数据集上与传统重叠社区发现算法COPRA和CPM相比,模块度值得到提高。With the surge in data volume,the relationship between data has become intricate and complicated,which brings challenges to network visualization.Through community discovery,highlighting the local clustering characteristics in the network can improve the visualization effect,and the discovery of overlapping communities can be closer to the actual network structure.The Louvain algorithm with simple,efficient and fast execution speed is currently one of the most commonly used community discovery algorithms,but the discovery of overlapping communities is its shortcoming.To this end,the paper is based on the Louvain algorithm,combined with the fuzzy C-means clustering algorithm based on spectral mapping to improve the community discovery algorithm.The improved algorithm uses spectral mapping to map data nodes to Euclidean space,the degree of membership is used to calculate the degree to which a data node belongs to a certain cluster,which allows the same data to belong to multiple different classes,thereby realizing the discovery of overlapping community structures.Finally,based on the proposed algorithm,the FR model in the mainstream layout algorithm is used to visualize the network data.Using the modularity value as an evaluation indicator,the experimental results show that the method proposed in the paper can find overlapping communities and can clearly show the community structure in the network,compared with the traditional overlapping community discovery algorithms COPRA and CPM on the classic data set,modularity values is improved.

关 键 词：社区发现 Louvain算法 模糊聚类方法 布局算法 图可视化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]