基于扰动的社交网络保护方法研究  

Research on perturbation-based privacy preserving method in social networks

作　　者：范国婷 杨颖[1] 孙刚[1] 赵佳[1] 

机构地区：[1]阜阳师范学院计算机与信息工程学院,安徽阜阳236037

出　　处：《阜阳师范学院学报（自然科学版）》2017年第4期55-60,共6页Journal of Fuyang Normal University(Natural Science)

基　　金：安徽省高校省级重点科研项目(KJ2017A332,KJ2016A549)资助

摘　　要：扰动技术是社交网络隐私保护的重要方法,本文提出了高斯随机扰动和贪心扰动两种扰动算法保护社交网络的权值,分别适用于动态和静态社交网络。高斯随机扰动可以简单有效地保护动态社交网络的权值隐私,贪心扰动算法将社交网络的边分类,可以在保护静态社交网络权值隐私的同时保证社交网络的最小生成树不变,提高社交网络数据的可用性。实验结果表明两种算法均能有效保护社交网络的权值安全,并且保持较高的数据可用性。Perturbation methods are crucial privacy-protecting approaches for social networks. The Gaussian perturbing randomly algorithm and greedy perturbation algorithm were put forward for the weights protection of the social networks. Gaussian perturbing randomly algorithm was simple to protect the weights of dynamic social networks. The edges were classified in the greedy perturbation algorithm, so that make minimum spanning tree same, meanwhile which can protect the weight privacy in static social networks. Experiment results show that two algorithms can effectively protect the weight information in social networks. The Gaussian perturbing randomly algorithm is fit for dynamic social networks while greedy perturbation algorithm can better protect static social networks.

关 键 词：社交网络 隐私保护 扰动 权值 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]