The Research On Privacy Preserving Data Publishing Based On Data Mining

Driven by benefits or by regulations that require certain data to be published，thereis a demand for the exchange and publication of data among various parties. Althoughdetailed person-specified data often contains sensitive information about individuals,and publishing such data directly violates individual privacy immediately. Therefore itis essential to have a research on the Privacy Preserving Data Publishing.The thesis studies the technique of PPDP based on data mining, and proposes aPPDP model based on data mining. In the model we obtain data publishing scenario byanalyzing attributes and Anonymilization operation, we also use data mining techniqueto guide the optimal selection of data publishing scenario. First of all, according to datapublication requirement we use association and correlation to analyze all the attributesin the data set and divide them into mutually disjoint attribute set. Then we use dataclassification to evaluate the risk of privacy disclosure and the usability of data to directthe anonymilization process.The experiment shows that the model can better satisfies the given demand fordata publishing and has the following advantages: dramatically reduce the risk ofprivacy disclosure while maintaining high data usability; the published data with lesserperturbation; it provides quantitative assessment of the usability of the published dataand the risk of privacy disclosure.