一种新的高维稠密数据隐私保护算法

A New Privacy Preserving Algorithm for High Dimensional Dense Data

针对隐私保护数据挖掘中的维数灾难问题,提出一种基于随机投影技术的隐私保护算法.该算法通过定义l投影扰动和Prevent-Ω数据集的概念,构造一种根据投影维数的不同,投影矩阵的稀疏度也相应变化的稀疏投影数据扰动,增加了数据的安全性.实验结果表明,在保护数据隐私的前提下,该算法能有效保证数据挖掘应用中的数据质量.

Aiming at the problem of the curse of dimensionality in privacy preserving data mining,we proposed a privacy preserving algorithm based on the technique of random projection.We defined the concept of l projection perturbation and Prevent-Ωdata set,and constructed a sparse projection data perturbation based on projection dimension of different projection matrix sparsity corresponding changes,which increased the security of data.The experimental results show that this method can effectively guarantee the quality of data in data mining applications under the premise of protecting data privacy.