An Enhanced Privacy-Preserving Method Based On K-automorphism Mode

In recent years,the rapid developments of Internet,data storage and computer technology make it possible to collect and analyze network information conveniently,and the process becomes more complete and accurate.However,the data publishing processes,which aim to information sharing,data mining,and knowledge discovery in database(KDD)and so on,are often accompanied by the leakage risk of sensitive privacy information.So,the research of privacy preserving in the data publishing(PPDP)is presented,aiming to improve the security of those released data security by losing some information in raw data appropriately,under the premise of guaranteeing the data utility,and then to provide a very good trade-off between privacy preserving and data utility.Based on the requirement of strengthening the privacy protection,this paper makes deep and detailed study on the anonymity techniques for privacy preserving data publishing in the case of ensuring the strong information utility.Firstly,this paper makes deep study on the privacy preserving models,which are based on anonymity technology.According to defects of these classical anonymity models,we introduce a privacy preserving algorithm named k-s,which is based on k-automorphism model.First,we should turn the original graph into a k-automorphism graph according to the frequent sub-graph we have chosen,and then it still have a problem that some sensitive information about link may be leaked out,so we should use LinkInfo Detection algorithm to check it out,we can guarantee the possibility of these information leaked out no more than 1/k through some operations of adding or deleting edges.The k-s algorithm makes less change about the structure of the original graph than the others,so it can help protecting the integrity of the original data,and also with better results on privacy-preserving.Finally,we conduct experiments with the help of public data sets,and compared with other algorithms in terms of many aspects,such as anonymity quality,anonymity cost,information loss and so no.Experimental results show that this approach can meet better privacy requirements and keep the information loss low,and it is a better privacy-preserving method.