Study On Privacy Protection Model

The modern society is known as the information society, as a result, the governmentdepartment, medical institution and other agencies are able to obtain authorization tostore and release data, so other organizations or individuals can gain all kinds of usefulinformation which can be applied in national defense construction, disease preventionor business strategy development, etc. But it can be bring about that personal privacy isexposed to the public. Therefore, data should be anonymous processed before data isreleased in order to protect the privacy of personal information. In the privacyprotection field, k-Anonymity model is widespread concern with its good practicalvalue, this paper research its advantages and disadvantages from technical Angle, andaccording to the defects of improved model, thus balance data privacy and availability.The main tasks of the paper are listed below:①Studying all kinds of existing privacy protection technology, this paper deeplyanalyze the k-Anonymity privacy protection model. Analyzed theoretically for a varietyof attacks to clear the way of corresponding leak way, and classify some of the classicalgorithm so as to equip the many problem areas of the algorithm, in order to determinethe advantages and disadvantages of the problem areas in existing solutions.②Aiming at the problem field of multidimensional sensitive attribute, this paperfrom the instance to analyze the existence of leak points, and based on the k-anonymitywith clustering features for multidimensional sensitive attribute similarity markers,finally propose a general anonymous model named Multi-dimensional sensitivek-anonymity privacy protection model which can give consideration tomultidimensional sensitivity, privacy, information accuracy and algorithm timecomplexity. Then, it gives details of the model implementation process, and from thetheoretically determine the feasibility of the algorithm.③According to the research and analysis about the re-publication of incrementaldatasets problem field, this paper puts forward the strategy which based on theMulti-dimensional sensitive k-anonymity privacy protection model to solve the problemdomain, and makes the model can give consideration to multidimensional sensitivity,re-publication of incremental datasets, privacy, information accuracy and algorithmtime complexity, etc. Thus, it better balance data privacy and availability.④Adult database from machine learning center of University of California is facto standard test sets in privacy protection area, the paper use it for experimental analysis,and obtains the performance situation about Multi-dimensional sensitive k-anonymitymodel and re-published strategy of incremental datasets, evaluation including algorithmtime complexity analysis, precision measurement analysis and privacy leak analysis.The experiments prove that algorithm time complexity and sensitive attribute dimensionare disproportionate, and the re-published strategy of incremental datasets can preventredundant calculation and leak from excessive change of equivalent group andeffectively protect the privacy information.