Research On K-Anonymity For Privacy Preserving

As an effective way for information exchange, data publishing provides convenience for data exchange and data sharing. But with the rapid development of data research, sensitive information disclosure is becoming a serious problem in data publishing. So privacy protection has become a new hotspot in data security research. The common way to protect privacy is to use K-anonymity in data publishing. This article will analysis the current research situation of K-anonymity and brings out a new model, which is is studied to solve the problem of sensitive information disclosure after anonymity, as follows:Firstly, on the basis of former research and existent solutions, (K, L)-anonymity is brought out from K-anonymity. This model will effectively eliminate attribute disclosure existing in K-anonymity. In order to guarantee the security in data publishing, Generation will be adapted and full domain generalization algorithm is proposed. Generation will provide an information loss formula for the data to be published. In addition, the correctness of the algorithm is proved and analyzed through examples.Secondly, clustering is applied to (L,K)-anonymity model, which translate K- clustering into K-anonymity and at last into (L,K) anonymity model. There is laboratory testimony and time analysis in the research.Clustering generation is proposed through the application of clustering distance measurement to (L,K) anonymity model, and there is also examples testimony and time complexity analysis.Finally, we draw a conclusion from the (L,K) anonymity experiment. The data from full domain generation algorithm can meet (L,K) anonymity's need, but a lot of information is lost. While clustering can keep the full informationbut it needs more times.