Reaserch On Differential Privacy Protection Based On Complex Network

The current social development is in a large data age.Information technology applications have changed a lot,massive data can be easily stored.However,the data set usually contain a lot of personal privacy content,the data through the data mining method can be very rich and valuable.Under the premise of protecting personal privacy,how to publish these data so that it can provide good data availability for data analysis and diggers has always been one of the key contents of privacy research.The concept of differential privacy is based on the assumption of the strongest background of the attacker,and the strict and quantifiable definition of privacy protection provide the strongest guarantee for privacy protection.Based on the differential privacy protection model,this paper studies the data release of the complex networks under non-interactive,and aims to publish complete data sets that satisfy the definition of differential privacy and have good usability.In the aspect of algorithm research,the main research contents of this paper is to propose two differential privacy data distribution algorithms,which are represent of small network model,based on the importance of nodes and the four-tree different privacy data released algorithm based on community partition are proposed.Then,the validity of the two algorithms proposed in this paper is illustrated by theoretical proof and experimental comparison.In the experimental research,the main research content is the complex network structure of the survivability of the analysis,this chapter through two improved differential privacy algorithm to protect the complex network model for network attacks,after a variety of attacks by attacking the network data fault tolerance and whether the attack capability is enhanced,and the attack strategy is improved to increase the diversity of attack types.