摘要
避免数据库中隐私数据的泄露,数据匿名化是实现隐私保护的一个有效手段。为了防止在多次发布中,链接攻击、背景知识攻击和同质攻击等引起的推理泄露,提出一种支持数据修改的隐私保护策略——m-inclusion规则。该规则同时运用了聚类算法,减少了信息的损失度,提高数据的实用性。与经典的m-invariance规则相比,更符合动态数据发布需求,不仅保护添加、删除时的隐私安全,还保证数据修改发布中的隐私安全。
To prevent data disclosure, data anonymity is an effective method to implement privacy protection. In order to prevent reasoning leak caused by linking attack, background knowledge attack and homogeneity attack etc. In multiple release, this paper puts forward a anonymous strategy to support data modification, called as m-inclusion principle. It also adopts clustering algo- rithm, which reduces the loss degree of information and improves the data practicality. Compared with m-invariance principle, m- inclusion more accords with dynamic data release demand. And it protects the privacy security flexibly.
出处
《计算机与现代化》
2013年第8期175-178,183,共5页
Computer and Modernization
基金
国家自然科学基金资助项目(61170221)
关键词
隐私保护
匿名化
数据重发布
数据修改
聚类
privacy protection
anonymous
data re-publication
data modification
clustering
