摘要
【目的】如何迅速且准确地识别关键谣言节点已成为当前研究的一项重要挑战。现有研究通常使用基于网络结构的中心性方法或基于节点特征的机器学习方法进行关键节点识别,然而这些方法大多基于信息传播的思路,忽视了谣言自身独特的传播方式。中心性方法未能充分反映节点在谣言传播中的实际重要性,即高中心性的节点未必在谣言传播中起关键作用;而机器学习方法则往往忽视谣言传播的结构信息。【方法】针对以上问题,提出了一种综合考虑谣言传播结构、信息传播结构以及用户属性信息的RumorGFAN模型,用于识别谣言传播过程中的关键谣言节点。此外,本研究考虑了谣言传播中的个人行为差异,即在接收到谣言后,一部分个人选择传播,而另一部分则选择不传播,采用了更符合谣言传播的易感-暴露-感染-康复-易感(SEIRS)模型,并提出了一种新的计算方法,以更准确地评估节点的影响力。【结果】在4个不同规模的真实数据集上的实验结果表明,该策略能够更准确有效地识别在线社交网络中的关键谣言节点。
【Purposes】How to quickly and accurately identify key rumor nodes has become an important challenge in current research.Existing studies usually use centrality methods based on network structure or machine learning methods based on node features to identify key nodes.However,most of these methods are based on the idea of information spreading,ignoring the unique way of rumor propagation.Centrality methods fail to fully reflect the actual importance of nodes in rumor propagation,that means nodes with high centrality may not play key roles in rumor propagation,while machine learning methods tend to ignore the structural information of rumor propagation.【Methods】To address the above problems,in this paper,a RumorGFAN model that integrates the rumor propaga-tion structure,information propagation structure,and user attribute information was proposed for identifying key rumor nodes in rumor propagation process.In addition,individual behavioral differences in rumor propagation were considered,after receiving a rumor,some individuals may may to spread it while others may not,Besides,the susceptible-exposed-infected-recovered-susceptible(SEIRS)model that is more consistent with rumor propagation was adopted,and a new computational method was proposed to assess the influence of nodes more accurately.【Results】Experimental results on four real datasets of different sizes show that this strategy is able to identify key rumor nodes in online social networks more accurately and efficiently.
作者
李琪
许晓雅
王莉
LI Qi;XU Xiaoya;WANG Li(College of Data Science,Taiyuan University of Technology,Jinzhong,Shanxi,China)
出处
《太原理工大学学报》
北大核心
2025年第5期887-896,共10页
Journal of Taiyuan University of Technology
基金
国家自然科学基金区域创新发展联合基金(U22A20167)
国家重点研发计划(2021YFB3300503)。
关键词
社交网络
关键谣言节点
节点识别
谣言传播
social network
key rumor nodes
node identification
rumor propagation
