基于数据特征的电力数据隐私保护模型研究

Research on Power Data Privacy Protection Model Based on Data Features

为避免发生隐私泄露和数据缺失,研究了基于数据特征的电力数据隐私保护模型。依据不同的数据属性划分原始数据集,形成特征集和候选集后,采用基于最大信息系数的特征分类模型,分析形成的两种数据集,获取最高的相关性形成隐私数据集;通过差分隐私的数据匿名隐私保护模型,利用差分隐私技术获取隐私保护匿名数据集,完成数据隐私保护。测试结果表明:模型在合理的隐私保护预算范围内,运算性能良好,保护后数据记录连接值低于0.23,可较大程度保证数据的隐私性和可用性,降低数据损失率,应用性较好。

In order to avoid the occurrence of privacy leakage and data loss,a power data privacy protection model based on data characteristics was studied.Divide the original data set according to different data attributes,form the feature set and the candidate set,and use the feature classification model based on the maximum information coefficient to analyze the two data sets formed to obtain the highest correlation to form privacy data set:through the data anonymity privacy protection model of differential privacy,use the differential privacy technology to obtain the privacy protection anonymous data set so as to complete data privacy protection.The test results show that the model is within a reasonable privacy protection budget range and has good operational performance.The record linkages value of the protected data are lower than 0.23,respectively,therefore,this idea can ensure the privacy and availability of the data to a greater extent,and reduce the data loss rate,and has good applicability.