欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
K-匿名化（K-anonymization）是数据发布环境下保护数据隐私的一种重要方法.目前,K-匿名化方法主要针对单一约束条件进行处理,而实际应用中涉及到大量的多约束条件,使K-匿名化问题更加复杂.如果简单地将单一约束K-匿名化方法应用到多约束情况,会造成大量的信息损失及过低的处理效率.根据多约束之间的关系,通过继承Classfly算法的元组概括过滤思想,提出多约束K-匿名化方法Classfly＋及相应的3种算法,包括朴素算法、完全IndepCSet算法和部分IndepCSet的Classfly＋算法.实验结果显示,Classfly＋能够很好地降低多约束K-匿名化的信息损失,改善匿名化处理的效率.
英文摘要：
K-Anonymization is an important approach to protect data privacy in data publishing scenario. Existing approaches mainly consider data processing with single constraint. There exist multiple constraints cases in the real applications, which makes the K-anonymization more complex. Simply applying the approaches with single constraint to the problem of multiple constraints may cause high information loss and low efficiency. Based on the idea of Classfly, a family of multiple constraints supported K-anonymization approaches named Classfly^＋ are proposed according to the features of mutiple constraints. Three K-anonymization approaches are proposed, which are naive approach, complete IndepCSet, and partial IndepCSet Classfly^＋ approaches. Experimental results show that Classfly^＋ can decrease the information loss and improve efficiency of k-anonymization.
同期刊论文项目
同项目期刊论文
位置：