Research On The Classification Method Of Privacy Protection Under Cyberspace

With the fast development of Internet,traditional industries and information technology have been merged together rapidly.For the convenience and speediness of"Internet plus",more and more human's daily activities are taken form "Offline" to"Online".Meanwhile,data generated by the activities are been recorded,which form a great fortune,"big data".However,when humans are enjoying the quick and convenient life which Internet brings to them,they also endure the risk of privacy information disclosure.In recent years,lots of privacy disclosure accidents have emerged continually,especially the network security accident introduced by "PRISM",which pushes privacy preserve problem to the cusp of public opinion again.Big data has the characters of variety,big volume,fast velocity and low density of value.If privacy data is protected at the same degree,it will waste the computing resource.In addition,over protected is not facilitate for the development of Internet and data mining.Therefore,the hierarchical privacy preserve strategy must be a future development orientation.According to the development experience of network security techniques,the entities' different demands of privacy preserve degree,and the cost of computing and storage,the hierarchical privacy preserve tactics will be a tendency in the future.The common data should be protected in a modest way,while high sensitive data should be protected carefully.The hierarchical privacy preserve way could avoid the imbalance of treating all kinds of data in the same way to protect users' privacy data in a more reasonable way.Besides,when new techniques occur,it could be a benchmark to evaluate the performance of the new techniques.It is also worked as a comparison standard to appraise the advantages and disadvantages of all kinds of techniques.The main works of this thesis are as follows:(1)The fundamental knowledge of privacy preserve is been reviewed by using an analysis tool named CiteSpace.This thesis is not only reviewing two kinds of privacy preserve techniques named k-anonymization and differential privacy,but also introducing the application of those techniques in graph structure privacy preserve.(2)Network degree centrality and network structure entropy are been picked as two kinds of privacy metrics index under network structure.Degree distribution theory in complex network is used to do privacy preserve research.By measuring and analyzing the value of four kinds of network basic characters named network average degree,network centrality,network entropy,and average clustering coefficient in five kinds of network mechanism model named star network,k-neighbour network,random network,small-world network,and scale-free network,we draw the conclusion that network centrality and network entropy could be the metrics index of network privacy.(3)A privacy preserve grading scheme based on degree distribution theory is been designed.This method makes use of network centrality and network entropy to quantify the privacy preserve degree.Assuming that k-neighbour network achieved privacy preserve entirely,it will be as a benchmark network structure.Together with the privacy evaluation model in privacy preserve international standards,and the percentage that all kinds of different network structures about the benchmark network structure,we define 6 different privacy preserve levels.(4)A pricing mechanism of differential privacy preserve service based on VCG mechanism is been proposed.Differential privacy is a strictly mathematical defined privacy preserve model that quantifies the degree of privacy preserve based on probability theory,which will make a solid foundation for privacy preserve levels.In the new scheme,we propose a hierarchical privacy preserve service model under the hypothesis that there have been some privacy preserve levels already.The model consists of VCG mechanism and optimal matching theory,which makes a reasonable price for every rank of services to adjust users' choice.Corresponding to the privacy evaluation model in privacy preserve international standard,the scheme makes price for 6 different privacy preserve levels.The pricing mechanism in the model can make a suitable price for every level,which is not only realizing the hierarchical privacy preserve scheme,but also optimizing the arrangement of social resources.