A Method Of Personalized Anonymity For Multimensional Privacy Preservation

The traditional ways of sensitive data usually use under the condition of a single sensitive property data, but the data often involve multiple sensitive properties in practical data applications. At present, the method of release multiple sensitive properties data have been an important research direction. But because of personal privacy protection on different data requirements and levels, the algorithms of reunification anonymous process used to a bit protection or short of protection. Therefore, the research of based personalized anonymity for multidimensional privacies preservation algorithms has become a new development of database security, and advance a new challenge to privacies preservation.At the beginning of this paper analysis the major existing privacy methods, and research the characteristics that existing multidimensional sensitive property data issued in privacy protection technology. According to the problems that the anonymous single sensitive property tables merge multidimensional sensitive property table cause redundant information, the thesis put forward a multidimensional sensitive property privacy protection technology that based on the least information loss strategy to improve anonymous model, and then bring up the matter of anonymous problem for the specific constraints of the gathering problem. Improve traditional cartesian algorithm applies to the loss of information when merge multidimensional sensitive property table. Secondly, because of the multidimensional sensitive property data technology in the process of anonymous don't consider the case of a single individual corresponding several records, which led to the problem that the same people loss the information of multidimensional sensitive property.This paper propose a new algorithm that based on the maintain identity of the multidimensional sensitive property, which adopt the connect of information loss and combine the approach (k, l)-anonymous data dissemination. Avoid overage generalized problems when a single individual corresponding records. At last the paper analysis and research the related development trend of individual anonymous technology and comprehensive the two characteristics of both the data security and the data effectiveness that based on the personal anonymous requirement of the multidimensional sensitive property application. By adopted more flexible strategy that the dynamic control technology of the multidimensional sensitive semantic categories tree, the paper make a resolution for privacy information divulges model when multidimensional sensitive property publish data of relational database, and gives the relevant algorithm.The paper use simulation experiment and use the Matlab for data processing, which specify a detailed result that a multidimensional sensitive property method in the information of redundancy and connect of information loss with individual privacy requirement. The method proposed in the paper that can overcome the shortcoming of redundancy information and better protect personal data sets for accuracy by the data in experiment and the results of analysis and synthesis. By simulation experiment results prove the fact that the method proposed can meet individual privacy for the minimum amount of generalizations and retain the maximum information in original data, which to improve the accuracy of data mining results.