基于聚类的数字图书馆用户隐私保护方法  

User privacy protection method of digital library based on clustering

作　　者：蔡晓峰[1] Cai Xiaofeng(Jiangsu Economic and Trade Vocational and Technical College,Nanjing 211168,China)

机构地区：[1]江苏经贸职业技术学院,南京211168

出　　处：《电子测量技术》2020年第2期123-127,共5页Electronic Measurement Technology

摘　　要：为提高数字图书馆用户隐私数据保护性能,提出了一种基于聚类的用户数据隐私保护方法。首先,利用数字图书馆和用户之间的关系数据构建网络模型。然后,利用综合特征距离对模型中节点进行聚类,将原始节点集合转化为至少包含k个节点的超点。最后,对超点中的节点属性信息进行概化处理,降低攻击者获取用户隐私数据的概率。实验结果表明,该方法的属性信息损失较低,匿名发布精度较高,且运算量较小,适合应用于数字图书馆用户隐私数据保护场合。To improve the performance of user privacy data protection in digital libraries, a clustering-based privacy protection method for user data is proposed. Firstly, the network model is constructed by using the relationship data between digital library and users. Then, the nodes in the model are clustered by using comprehensive feature distance, and the original set of nodes is transformed into a super-point containing at least k nodes. Finally, the attribute information of nodes in the super-point is generalized to reduce the probability of attackers obtaining user privacy data. The experimental results show that this method has low loss of attribute information, high accuracy of anonymous publishing, and low computational complexity. It is suitable for the application of privacy data protection in digital libraries.

关 键 词：数字图书馆 隐私保护 聚类 属性概化 

分 类 号：TN391[电子电信—物理电子学] TP309[自动化与计算机技术—计算机系统结构]