Font Size: a A A

The Research On Structured Query Language Satisfying Differential Privacy

Posted on:2021-04-11Degree:MasterType:Thesis
Country:ChinaCandidate:C Z ChangFull Text:PDF
GTID:2428330602989077Subject:Engineering
Abstract/Summary:PDF Full Text Request
With the continuous development of the information age,the effective information in the mining data has become more and more popular,resulting in more and more attention to privacy protection in the process of mining information.As a new type of privacy protection model,differential privacy provides strict mathematical definitions of privacy protection and provides quantitative evaluation methods,so that the degree of privacy protection provided by data sets under different parameter processihg is comparable.Therefore,after the differential privacy theory was proposed,it was quickly recognized by the industry and gradually became a research hotspot in the field of privacy protection.At present,differential privacy is relatively mature in theory,but it is still limited in practical applications.Structured Query Language(Structured Query Language,SQL)is a kind of non-procedural data query language widely recognized and popularized in the real world,with high flexibility and powerful features.While SQL has great flexibility,it also increases the difficulty of adding differential privacy protection,resulting in insufficient support for differential privacy by SQL queries.In order to expand the use of differential privacy in practical applications and improve the support of SQL queries for differential privacy,this paper proposes a method for adding differential privacy protection to SQL queries.This paper analyzes the differential privacy protection methods of commonly used aggregate functions in SQL,and divides the aggregate functions in SQL into two categories.Since the function value of the first type of aggregate function is affected by the size of the attribute value in the data set,the global sensitivity of the first type of aggregate function is high,resulting in poor usability of the results after privacy protection.In order to improve the usability of published results,this paper proposes a method to add differential privacy protection to SQL queries through local sensitivity.In addition,for the second type of aggregate function,a method to provide differential privacy protection for SQL queries under correlative datasets is proposed.From the perspective of relational algebra,this method restricts SQL through relational algebra,and skillfully solves the problems caused by the flexible SQL query structure.In addition,the correlative data set is a highly sensitive data set.When one of the records is changed,multiple records may be changed.In view of this situation,this paper uses correlation coefficients to measure the correlation between data,and fully considers the impact of correlation on privacy protection results in the process of differential privacy protection,and strives to improve the availability of published results under the premise of satisfying differential privacy protection.
Keywords/Search Tags:correlative datasets, structured query language, privacy protection, differential privacy
PDF Full Text Request
Related items