基于改进贝叶斯网络的高维数据本地差分隐私方法

Method of Local Differential Privacy Method for High-Dimensional Data Based on Improved Bayesian Network

文章提出一种基于改进贝叶斯网络的高维数据本地差分隐私方法,首先通过数据源差分隐私保护算法对用户端数据集进行扰动,生成扰动数据集,保护本地原始数据集隐私;然后通过改进的贝叶斯网络将高维数据集降维为多个低维属性集合;最后合成新数据集,使用人工蜂群算法对贝叶斯网络结构进一步改进。实验结果表明,该方法在数据实用性方面具有优势,且得到的贝叶斯网络收敛性更好。

In this paper,a local differential privacy method for high-dimensional data based on improved Bayesian network was proposed.By using the differential privacy protection algorithm of data source,the client data set was disturbed to generate the disturbed data set,so that the privacy of the local original data set was protected,and the privacy security of users was fundamentally protected.Then the high-dimensional data set was reduced to several low-dimensional attribute sets by the improved Bayesian network,and the new data set was finally synthesized,and the artificial bee colony algorithm was used to further improve the construction of Bayesian network structure.Finally,the experimental results show that the research method in this paper has advantages in data practicability,and the Bayesian network structure achieved better convergence.