首页|基于多维桶分组技术改进算法对电子病历隐私信息研究

基于多维桶分组技术改进算法对电子病历隐私信息研究

扫码查看
多维桶分组技术给出的三种算法在敏感属性数量增多时,发布数据隐匿率增大,且该算法对准标识符属性没有泛化处理,易遭受背景知识攻击。针对此问题,提出改进的最大桶优先算法。该算法依据敏感属性关联性来对敏感属性表进行划分,每个敏感属性子表采用最大选择度优先的方法实现匿名化处理,并增加对准标识符属性泛化处理;增加对每个等价类中的复合敏感属性向量随机交换来抵御关联攻击。实验结果表明,改进算法能够在使得发布数据保持一定数据可用性基础上更好地保护用户的敏感属性信息。
PRIVACY INFORMATION OF ELECTRONIC MEDICAL RECORD BASED ON IMPROVED ALGORITHM OF MULTI-DIMENSIONAL BUCKET GROUPING TECHNOLOGY
The three algorithms based on multi-dimensional barrel grouping technology increase the hidden rate of published data because of more diversities between sensitive properties when the number of sensitive attributes increases.Moreover,it is easy to be attacked by background knowledge when the algorithm is not generalized to the identifier attribute.Aiming at such problems,this paper proposes an improved maximum bucket priority efficient algorithm.The correlation of sensitive attributes was analyzed in this algorithm to reduce the dimension of sensitive attributes.For each sensitive attribute sub-table,the maximum preference degree was used to realize the anonymization processing,and increase the generalization of the identification attribute generalization.A random exchange of complex sensitive attribute vectors in each equivalence class was increased to ward off the associated attacks.The experimental results show that the improved algorithm can protect the sensitive attribute information of users better on the basis of maintaining data availa-bility.

Privacy protectionBackground knowledge attackMulti-sensitive attributeCorrelation

张付霞

展开 >

贵州民族大学民族文化与认知科学学院 贵州 贵阳 550025

隐私保护 背景知识攻击 多敏感属性 相关性

贵州大学研究生创新基金项目

校研理工2015017

2024

计算机应用与软件
上海市计算技术研究所 上海计算机软件技术开发中心

计算机应用与软件

CSTPCD北大核心
影响因子:0.615
ISSN:1000-386X
年,卷(期):2024.41(2)
  • 4