Research On Privacy-Preservation Anonymity Method For Data Sharing

With the rapid development of network information technology and data miningtechnology can help people solve problems more and more outstanding, data sharing as asource exchange way that no doubt has the very impotent position for various purposes ofresearch works for government departments,commercial organizations, research institutionsand so on.So how to prevent personal privacy from being betray when data sharing providesconvenient to users,which is a significent problem. k-anonymity pattern is an effective methodto keep privacy data safeing,however,because of the increasingly high demand for datasecurity,in recent years,for the purpose of enhanced data security to improving and optimizingthis anonymous data privacy protection methods is very urgent.Face to current data anonymity privacy protection models less consider how to deal withsome sensitive attribute values depend on some values of others attribute column,and relatedwith demands of the diversity and sensitivity in different applications，in this thesis，base onsensitive divide and priority cluster related attribute value method, we put forword an(d,α)k-anonymity model.Through defining some constraint rules,the disclosure of privacyinformations that because the relationship between sensitive attribute values and other valuescan be controlled,it can also reduce the useful information has been modified.In this thesis,toreduce the information loss and to solve the shortcoming which processing the categoricaldata exist in the current anonymous algorithms,we introduce the cluster algorithm into the(d,α)k-anonymity model,by calculating the distance between individuals to ensure theequivalence classes as similar as possible,in order to enhanced the flexibility and accuracycompared with the traditional full domain generation algorithm.Finally,the simulation expriment results show that the method is effective to protectpersonal privacy in the context of information,it is able to effectively reduce the loss ofinformation result from data anonymization process.