一种基于随机投影的加权社会网络隐私保护方法

Privacy Preserving Method Based on Random Projection for Weighted Social Networks

针对加权社会网络的发布,提出了一种基于随机投影的隐私保护方法——向量集随机投影,该方法通过对加权社会网络的结构和边权重进行干扰实现敏感信息的隐私保护。通过对加权社会网络进行分割,得到节点数相同的若干个子网络;依据边空间理论,采用由边信息构建的向量描述子网络,构建加权社会网络的向量集作为发布模型;利用随机投影技术对原始向量集进行降维操作得到目标向量集;依据目标向量集构建加权社会网络的发布集。实验结果表明,向量集随机投影方法能够在确保隐私信息安全的同时仍然保护社会网络分析所需要的某些结构特征。

A privacy preserving method based on random projection namely vectors set random projection was put forward on the publication of weighted social networks.The method protects sensitive information security through perturbing network structures and edge weights.It partitions weighted social networks into multiple sub-networks with the same number of nodes.Based on the theory of edge space,it describes the sub-networks by vectors consisted of edges information and constructs vector set of weighted social networks as the released model.It uses random projection technology for dimension reduction and maps the original vector set into the targeted vector set.It constructs the released weighted social networks based on the targeted vector set.The experimental results demonstrate that the vector set random projection method can ensure privacy information security and protect some structure characteristics of the social network analysis.