Research On Privacy Preserving Mechanism For Multi-sensitive Attributes Based On Time Sequence

With the rapid development of computer and network technology,application and sharing of database are more and more widely used,that will bring about privacy information leakage of important data.Many researchers have conducted the exploration relevantly,which has formed a series of anonymous rules to protect the data private.However,the existing data publishing mostly concentrates on the one tine and usually considers the sensitive attributes as equivalent.Most of the practical applications may contain multiple releasing versions,meanwhile,the sensitive attributes also exist the reasoning associations.Therefore,secure releasing of private data not only is a problem to be solved,but also is the hots pot in the direction of researching.In this paper,research work is as follows:First,the current privacy protection technology seldom considers the defects of the dynamic data model,a privacy reasoning association model is introduced and the association between the versions of multiple releases are also established.In the process of researching the data publication,sensitive attributes are associating with the presence of reasoning.The thesis systematically proposes the identification and classification thought to multi-sensitive attribute.Combining with the constraint of database paradigm,this paper gives a concreted implementation mechanisms and methods about this inference model.Secondly,the paper uses the Markov chain method to describe the properties of the releasing behaviors,and combines with the characteristics of state-independent features.During the process of data releasing,it can construct a time sequence and operating behaviour queue,forming an associative model and determined method.The impactive factors are calculated about each attributes based on the rules of association reasoning,and the local safety threshold is computed as well.Finally,based on the existing privacy protection,this paper extends the security policy about the time constraint,constructing the metric models,security formulas and the global security.During the process of data publishing,the algorithms are designed to determine whether the privacy is leakage or not,and it adopts an anonymous strategy to prevent the information disclosure.The experiment is verified that the algorithms not only can effectively reduce the loss rate of attributes during releasing information which can improve the running time of the system,but also enhance the security and authenticity of the data.