摘要
基于启发式的隐私保护关联规则挖掘算法常通过删除项或增加项改变规则的支持度,现有的通过删除项的隐私保护关联规则挖掘算法设计过程中通常忽略了兴趣度和规则的左件,导致对非敏感规则的支持度和数据可用性影响很大。针对上述不足,在算法设计过程中引入了兴趣度和逐步移项的思想,通过对敏感规则的左右件选择性地适当处理,不仅成功隐藏了指定隐私规则集,同时降低了对非敏感规则支持度的影响,提高了数据的可用性。理论和实验结果表明i,f-then算法具有较好的隐私性和高效性。
In order to change the support of rules, the heuristic approaches are usually realized by deleting an item or inserting item. Existing privacy preserving association rule mining algorithms by removing item usually neglected interest measure and the left hand of rules, which affect non-restrictive rule support and availabity ofdata negatively. Focusing on the shortcoming, an effective privacypre- serving algorithm is proposed. Trough bringing interest measure and proper item （left or right part） deal, reducing modifying ratio, a greater degree of hiding is made while a less degree impact for the non-sensitive rules can be made too. Theoretical analysis and experimental results show that the algorithm of if-then is highly efficient and has got good privacy.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第6期2124-2128,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(50674086)
关键词
隐私保护
关联规则
数据挖掘
兴趣度
敏感规则
privacy preserving
association rule
data mining
interest measure
sensitive rules
