基于粒子群算法的社会网络k-度匿名图修改方法  

k-degree anonymous graph modification method for social networks based on particle swarm optimization algorithm

作　　者：李晓晔[1] 王小进 LI Xiao-ye;WANG Xiao-jin(College of Computer and Control Engineering,Qiqihar University,Heilongjiang Qiqihar 161006,China)

机构地区：[1]齐齐哈尔大学计算机与控制工程学院,黑龙江齐齐哈尔161006

出　　处：《齐齐哈尔大学学报(自然科学版)》2025年第1期27-35,共9页Journal of Qiqihar University(Natural Science Edition)

基　　金：黑龙江省省属高等学校基本科研业务费科研项目(145209124)。

摘　　要：针对当前社会网络的匿名化隐私保护方法存在信息损失量大,忽略社会网络的结构等问题,提出一种保护社会网络社区结构的基于粒子群算法的k-度匿名方法。首先,使用贪婪算法对社会网络图的节点进行划分,得到节点欲达成k-度匿名所需增加的度数序列;其次,引入社区发现,减少图结构的损失;最后,基于粒子群算法对图进行边添加,满足k-度匿名。实验使用平均路径长度、平均聚类系数和传递性作为评价指标,在3个数据集上对提出的方法进行实验测试。结果表明,该方法能抵御度属性的攻击,较好地保护了网络图的社区结构,同时降低了图的信息损失量。To address the issues of significant information loss and the oversight of social network structures in current anonymization methods for privacy protection in social networks,a k-degree anonymity method based on particle swarm optimization is proposed to protect the community structure of social networks.Firstly,employs a greedy algorithm to partition the nodes of a social network graph,resulting in a sequence of degree increments required to achieve k-degree anonymity.Secondly incorporates community detection to minimize the loss of graph structure.Finally,uses particle swarm optimization to add edges to the graph,ensuring it meets k-degree anonymity.Experiments were conducted using the average path length,average clustering coefficient,and transitivity as evaluation metrics on three datasets to test the proposed method.The results demonstrate that this method not only defends against degree-based attacks but also effectively preserves the community structure of the network graph while reducing the amount of information loss.

关 键 词：社会网络 k-度匿名 粒子群算法 图修改 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]