Research On Rational Privacy Protection Model And Algorithm Based On Association Rule

Sharing and open of data is the key foundation for data analysis and application,and data publishing is an important wat to open and share data.However,the direct release of data faces the threat of privacy leakage.Particularly,the relevance between attributes of multi-source data exacerbates the risk of privacy leakage.Therefore,research on privacy-preserving for data publishing has become an urgent problem to be solved.Based on the correlation between attributes of structured data,this paper discusses the rational privacy protection model and scheme based on association rule,aiming at solving the problem of balance between privacy protection and data utility.Firstly,based on the anonymous privacy protection model,considering the correlation among data attributes,we proposed an optimization privacy protection model based on association rule to make up for the shortcoming of the existing anonymous privacy protection model.Secondly,a rational privacy protection model is proposed from game theory.It seeks to resolve conflict between privacy protection and data utility by game equilibrium analysis.The specific research work is as follows:(1)An anonymous privacy protection model based on association rule is proposed.The model establishes association rules among structured data attributes,and quantifies the correlation by mutual information.Based on p-sensitive k-anonymity model,the model set the threshold of privacy protection to avoid the weakening of utility caused by excessive data protection.Finally,experiment analysis verifies the effectiveness of privacy protection model.(2)An anonymous privacy protection model based on game theory is designed.Aiming at the confit between privacy protection and data utility,we regards privacy protection and data utility as two sides of the game to construct a complete information dynamic game model.The model realizes the design of the utility function according to the interaction between each participant.Base on this,the solution method of Nash equilibrium is given.Finally,it's applied to the t-closeness based on micro-aggregation privacy protection model.The experiment shows that the model can solve the problem of optimal parameter setting in anonymity model.(3)Based on the above research,a rational privacy protection scheme based on association rule is designed.Based on the classification of sensitive attributes,the scheme constructs an equivalence class partitioning strategy and proposes a rational privacy protection scheme for multi-dimensional data.The comparative analysis of related privacy protection scheme reflects the advantages of the scheme;and experiment shows that the scheme can achieve data of a certain degree of privacy protection clustering.