一种医疗数据发布匿名化模型  

An Anonymous Model of Medical Data Publishing

作　　者：冷建宇 郭永安[1] Leng Jianyu

机构地区：[1]南京邮电大学通信与信息工程学院,江苏南京210003

出　　处：《工业控制计算机》2021年第4期60-62,65,共4页Industrial Control Computer

基　　金：江苏省研究生科研与实践创新计划项目(SJCX19_0240)。

摘　　要：针对疾病这个敏感属性包含两重语义信息的特点,提出了一种(w,k,d)-匿名模型。该模型首先对疾病的敏感等级进行划分,计算每种疾病所在分级的权重值,限制每个等价类的平均权重值不大于给定的约束值w;其次,按照语义层次树对疾病进行划分,要求等价类的平均语义层次距离不小于给定的约束值d,最终实现对于疾病这个敏感属性的个性化保护。实验数据表明,尽管消耗了一些执行时间,但是却能更有效地阻止疾病属性被相似性攻击,保护病人隐私。According to the characteristic that the sensitive attribute of disease contains two semantic information,this paper proposes a(w,k,d)-anonymity model.Firstly,this model divides the sensitive grades of diseases,calculates the weight value of each disease grade,and limits the average weight value of each equivalent class to be no greater than the given constraint value w.Secondly,diseases are divided according to the semantic hierarchy tree,and the average semantic hierarchy distance of equivalent classes is required to be no less than the given constraint value d,so as to realize the personalized protection of the sensitive attribute of diseases.Experimental data show that although some execution time is sacrificed,it can prevent disease attributes from being attacked by similarity more effectively and protect patients′privacy.

关 键 词：隐私保护 (p α)-sensitive k-anonymity模型 分级 (w k d)-匿名模型 

分 类 号：R-05[医药卫生] TP309[自动化与计算机技术—计算机系统结构]