Abstract

Privacy-preserving data mining is used to safeguard sensitive information from unsanctioned disclosure. Privacy is an important issue in data publishing years because of the increasing ability to store personal data about users. A number of techniques such as bucketization, generalization have been proposed perform privacy-preserving data mining. Recent work has shown that generalization not support for high- dimensional data. Bucketization cannot prevent membership disclosure and does not apply for data that do not have a clear separation between quasi-identifying attributes and sensitive attributes. A new technique is introduced that is known as slicing, which partitions the data both horizontally and vertically. Slicing provides better data utility than generalization and can be used for membership disclosure protection. Slicing can handle high-dimensional data. Also slicing can be used for attribute disclosure protection and develop an efficient algorithm for computing the sliced data that obey the l-diversity requirement. Slicing is more effective than bucketization in workloads involving the sensitive attribute. Another advantage of slicing can be used to prevent membership disclosure.

Authors and Affiliations