Reasearch And Implementation On Differential Privacy Publishing Method Of Population Data

Population data is of great significance to the formulation and improvement social policies and promotion high-quality economic development.It is valuable for research but also highly sensitive.However,the current publishing methods of population data is hard to offer users query results with privacy and data availability.With the increasing attention of the society to the privacy protection problem,differential privacy has gradually become a focus of academic research due to its strict mathematical definition and quantifiable privacy leakage risk.Therefore,this paper will research the publishing method of population data under differential privacy protection.Firstly,this paper introduces the current main publishing methods and common attack types of population data,and analyzes the problems of traditional privacy protection technology in data release,then introduces the definition,property and release framework of differential privacy in detail.Secondly,this paper designs a publishing method HC-PRDP based on hierarchical control of population data under differential privacy protection.The HC-PRDP method uses blockchain and differential privacy technology to enhance the security and privacy of population data release.In the HC-PRDP method,the access control layer uses smart contracts to classify users,and the transaction is verified on the network and added to the blockchain,and then the user permissions are passed to the differential privacy layer.Differential privacy layer obtains privacy budgets according to the differences in user authority values,user reputation values,and data protection levels,and hierarchical queries are provided for users,so that different users can get publishing data with different availability.In addition,for common statistical query methods,the HC-PRDP method implements mean query,top-k query and histogram query that satisfy differential privacy.At last,the experiments prove the privacy of the HC-PRDP method and the availability of released data for different users.Finally,in order to solve the problem of low availability in released data caused by noise accumulation as the query range increases during the histogram publishing,this paper proposes the differential privacy histogram publishing optimization algorithm OBA-HP.The algorithm establishes the relationship between the added noise and the release error,by analyzing the error generated in the process of publishing histograms.And according to the property of the derivative,it obtains the formula between the distribution ratio of the privacy budget and the grouping result when the error in released data is the smallest.Then the privacy of the algorithm is proved through theoretical analysis.At last,it is proved that the OBA-HP algorithm can improve the accuracy of published data by comparative experiments.