Research On Data Publishing Methods Based On Differential Privacy

With the rapid development of Internet technology,our daily lives are gradually networked,and more and more personal information is published online.Every day,huge amounts of information flow between major Internet platforms.Various privacy information leakage problems have appeared in major online platforms.Attackers gain benefits by leveraging leaked private information.When our private information is leaked,we may be harassed in various ways,even involving personal safety.Although the industry has begun to attach importance to privacy protection and adopted corresponding protection measures,the expected results are not satisfactory.the technology of data publishing has been continuously researched and improved,and it has become a research focus of many scholars.By studying the traditional data publishing methods,it is found that the classification effect and accuracy are not ideal.To this end,this paper proposes a new data publishing method based on differential privacy.First,the original data set is generalized,second classify the generalized data set,third proceed with privacy budget allocation,finally,the data set is processed with noise to satisfy the differential privacy conditions.In the process of data classification,a classification tree formed by selecting segmentation points based on existing data information characteristics cannot meet requirements.This paper analyzes the data characteristics and proposes a new type of data information characteristics.In the process of privacy budget allocation,by studying several traditional privacy budget allocation methods,this paper proposes a new privacy budget allocation method,which can more reasonably allocate privacy budgets,while also avoiding waste of privacy budgets.Finally,in this paper,three data sets are selected for comparison with two data publishing methods.Using the control variable method,experiments with the number of classification layers and privacy budget as independent variables,through the experimental results,it is verified that the method proposed in this paper can improve the accuracy of data release while ensuring the reasonable allocation of privacy budget.