摘要
随着5G的发展,网络传播以前所未有的速度向各领域渗透.其中,影响力分析是研究网络信息传播机制的关键技术.传统的影响力分析算法主要通过选取具有最大传播特性的种子节点用于网络传播.但在种子节点选取上,其度量算法没有反映社交网络中的潜在信息.这将对影响力传播分析造成影响,同时一旦社交网络结构遭到破坏,网络的传播能力将会受到影响.针对这一问题,本文首先定义了一种基于属性的朋友亲密度度量关系,量化网络中各用户的影响力;其次,提出了一种属性和亲密度兼顾的影响力算法,该算法综合考虑了网络的结构、属性和亲密度三者之间的关系,选择出具有高影响力及抗攻击性的种子节点,提高网络传播能力及抗攻击能力;最后,通过真实网络环境下的实验,验证出相比现有的度量算法,本文提出的算法在种子节点数目较大时具有更好的传播特性,并且在隐私高风险状态下的社交网络中,该算法的受攻击影响程度稳定在5%-10%左右,影响程度最低,具有较好的抗攻击性.
With the development of 5 G,network propagation penetrates into all fields at an unprecedented speed.Then influence analysis is the key technology to study the mechanism of network information transmission.In the traditional influence measurement algorithms,the nodes with maximum influence are explored as the seed nodes,using for network propagation.However,in the selection of seed nodes,this kind of metric algorithm cannot reflect the potential information in the social network,which will take some implications for the analysis of influence-communication.At the same time,once the social network structure is destroyed,the spread of network capacity will be affected.To solve this problem,firstly,an attribute-based friend closeness measurement relationship is defined in this paper,which can quantify the measure of influence;secondly,a wandering algorithm based on the relationship between attributes and affinity-density is proposed,which integrates the structure of the network and the relationship between attributes and affinity-density,then,the seed nodes with high influence and resistance to aggression can be selected;lastly,the experiments show that,compared with the traditional measurement algorithm,the algorithm proposed in this paper has better propagation characteristics when the number of seed nodes is large.Moreover,in the social network with high privacy,the influence degree of the algorithm is stable at about 5%-10%,and the influence degree is the lowest.Therefore,the algorithm has a better resistance to aggression.
作者
胡健宇
章静
许力
林力伟
HU Jian-yu;ZHANG Jing;XU Li;LIN Li-wei(School of Computer Science and Mathematics,Fujian University of Technology,Fuzhou 350118,China;Fujian Provincial Key Laboratory of Big Data Mining and Applications,Fujian University of Technology,Fuzhou 350118,China;Fujian Provincial Key Lab of Network Security&Cryptology,Fujian Normal University,Fuzhou 350007,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2022年第2期422-429,共8页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61902069,U1905211)资助。
关键词
社交网络
影响力最大化
亲密度
属性
抗攻击性
social network
influence maximization
affinity-density
attribute
anti-attacking
