软件开发数据库中隐私数据保护仿真

摘要

对软件开发数据库中隐私数据的保护,能够有效提高数据库安全性.对数据库隐私数据的保护,需要根据所求得原始矩阵的低阶近似矩阵,获取隐私数据特征的最佳投影,完成隐私数据的保护.传统方法通过关联规则方法实现数据清理,过除部分敏感项,但忽略了数据特征的最佳投影,导致隐私数据保护效果不理想.提出基于差分保护的软件开发数据库中隐私数据保护方法.采用指数机制将数据逐步精确化,利用差分隐私的序列组合性质对数据进行泛化保护,融合于主成分分析理论建立原始数据矩阵的协方差矩阵,得到原始矩阵的低阶近似矩阵,获取隐私数据特征的最佳投影,以此完成软件开发数据库中隐私数据保护.实验证明,所提方法下隐私数据安全性较高,可以大幅度地提升软件信息系统开发的质量.%The protection of privacy data in software development database can effectively improve the security of database.For the protection of privacy data in database,traditional methods use association rule to realize the data cleaning,but ignore the optimal projection of data characteristic,resulting in the unsatisfactory effect of privacy data protection.This article presents a method for protecting privacy data in software development database based on difference protection.The exponential mechanism is used to make data precise stage by stage.The properties of sequence combination of difference privacy are used for generalization protection of data.Integrated with the theory of principal component analysis,a covariance matrix of original data matrix is established,a low-order approximate matrix of original matrix is obtained,and the optimal projection of privacy data feature is achieved.Thus,the protection of privacy data in software development database is completed.Simulation results show that the proposed method has high security of privacy data,which can greatly improve the quality of information system development of software.