CurrentDevelopmentsofk-AnonymousDataReleasing.doc

Current Developments of k-Anonymous Data Releasing Jiuyong Li, Hua Wang2, Huidong Jin3, Jianming Yong4 1School of Computer and Information Science, University of South Australia, Mawson Lakes Adelaide, Australia, 5095 Email: jiuyong.li@.au 2Department of Mathematics and Computing, University of Southern Queensland, Toowoomba, Australia 4350 Email: hua.wang@.au 3National ICT Australia (NICTA), Canberra Lab, ACT, Australia, 0200, huidong.jin@.au. 4Department of Information Systems, University of Southern Queensland, Toowoomba, Australia, 4350 Email: yongj@.au Abstract Disclosure-control is a traditional statistical methodology for protecting privacy when data is released for analysis. Disclosure-control methods have enjoyed a revival in the data mining community, especially after the introduction of the k-anonymity model by Samarati and Sweeney. Algorithmic advances on k-anonymisation provide simple and effective approaches to protect private information of individuals via only releasing k-anonymous views of a data set. Thus, the k-anonymity model has gained increasing popularity. Recent research identifies some drawbacks of the k-anonymity model and presents enhanced k-anonymity models. This paper reviews problems of the k-anonymity model and its enhanced variants, and different methods for implementing k-anonymity. It compares the k-anonymity model with the secure multiparty computation-based privacy-preserving techniques in the data mining literature. The paper also discusses further development directions of the k-anonymous data releasing. Keywords: privacy preserving, data releasing, k-anonymity. Introduction Various organisations, such as hospitals, medical administrations and insurance companies, have collected a large amount of data over years. However, gold nuggets in these data are unlikely to be discovered if the data is locked in data custodians’ storage. A major risk of releasing data for public research is revealing the private information of i