摘要: 在区域医疗信息共享下,传统的匿名化隐私保护算法面对背景知识攻击时抵抗力较差。为此,提出一种敏感属性聚类匿名算法。利用敏感属性之间的关联进行微聚类,使等价组中敏感属性之间在相似性增大的同时存在差异性,从而较好地抵抗背景知识攻击,提高抗泄露风险能力。实验结果表明,该算法能减小数据信息表中的隐私泄露风险。
关键词:
区域医疗,
信息共享,
隐私保护,
信息聚类,
敏感属性,
背景知识
Abstract: In the regional medical information sharing situation, traditional anonymity privacy protection algorithm performs poor resistance when facing with background knowledge attack. So this paper introduces the anonymous sensitive attribute clustering algorithm. The algorithm uses the links between sensitive attributes to make micro-property clustering, it increases the similarity between the sensitive attributes in the equivalent group while remaining the difference between them. The algorithm can perform better resistance while facing background knowledge attack with improved ability to resistant the leakage risks. Experimental results show that the loss of privacy in the data sheet risk reduced substantially when using the algorithm.
Key words:
regional medical treatment,
information sharing,
privacy protection,
information clustering,
sensitive attribute,
background knowledge
中图分类号:
陈炜, 陈志刚, 邓小鸿, 黄伟琦. 抵抗背景知识攻击的电子病历隐私保护新算法[J]. 计算机工程, 2012, 38(11): 251-253.
CHEN Wei, CHEN Zhi-Gang, DENG Xiao-Hong, HUANG Wei-Qi. Novel Algorithm on Electronic Medical Record Privacy Protection Against Background Knowledge Attack[J]. Computer Engineering, 2012, 38(11): 251-253.