基于差分隐私的多源数据关联规则挖掘方法

Mining Method of Association Rules Based on Differential Privacy

随着大数据时代的到来,挖掘大数据的潜在价值越来越受到学术界和工业界的关注。但与此同时,由于互联网安全事件频发,用户越来越多地关注个人隐私数据的泄露问题,用户数据的安全问题成为阻碍大数据分析的首要问题之一。关于用户数据的安全性问题,现有研究更多地关注访问控制、密文检索和结果验证,虽然可以保证用户数据本身的安全性,但是无法挖掘出所保护数据的潜在价值。如何既能保护用户的数据安全又能挖掘数据的潜在价值,是亟需解决的关键问题之一。文中提出了一种基于差分隐私保护的关联规则挖掘方法,数据拥有者使用拉普拉斯机制和指数机制在数据发布的过程中对用户数据进行保护,数据分析者在差分隐私的FP-tree上进行关联规则挖掘。其中的安全性假设是:攻击者即使掌握了除攻击目标以外的所有元组数据信息的背景知识,仍旧无法获得攻击目标的信息,因此具有极高的安全性。所提方法是兼顾安全性、性能和准确性,以牺牲部分精确率为代价,大幅增加了用户数据的安全性和处理性能。实验结果表明,所提方法的精确性损失在可接受的范围内,性能优于已有算法的性能。

With the advent of the era big data,the potential value of mining big data has attracted more and more attention from academia and industry.However,at the same time,due to frequent Internet security incidents,users are increasingly concerned about the disclosure of personal privacy data,and user data security issues become one of the most important obstacles to big data analysis.With regard to the study of user data security,the existing researches more focus on access control,ciphertext retrieval and result verification.The above researches can guarantee the security of user data itself,but can not dig out the potential value of protected data.Therefore,how to protect the security and dig the potential value of the data in the meantime is one of the key issues that need to be addressed.This paper proposed an association rules mining method based on differential privacy protection.Data owners use Laplacian mechanism and exponential mechanism to protect user data during data release.Data analysis is associated with differential privacy FP-tree Rule mining.The experimental results show that the performance and accuracy of the proposed method are superior to the existing methods.