Slicing+: An Efficient Privacy Preserving Data Publishing

Abstract

Objectives: Privacy and accuracy are always trade off factors in the field of data publishing. Ideally both the factors are considered critical for data handling. Privacy loss and accuracy loss need to be maintained low as possible for an efficient data handling system. Authors have come up with various data publishing techniques aiming to achieve balance between these 2 factors. Generalization, Bucketization and Slicing are well known techniques among the list. Unfortunately they have their own limitation in handling privacy and accuracy. Generalization suffers in handling high dimensional data thus experiencing higher utility loss. Bucketization lacks data privacy where parting sensitive and quasi identifier attributes is a challenge. Slicing on the other hand though offers better privacy and accuracy, there is always scope to improve data correlation aiming in reducing utility loss. This paper explains a new technique called Slicing+ which handles privacy and accuracy factors effectively. This new slicing+ technique looks promising as it offers flexibility for data publisher to decide on how the data need to published. Data publisher can tune the Slicing+ technique to get data published with better privacy than accuracy or the other way. Algorithms for the two cases are derived and realized usingORANGE tool. This paper explains analysis done for the first bucket tuples. As an improvement aspect, similar analysis can be done for other buckets and all the bucket tuples merged and reconstructed for complete analysis. This analysis is applied in the medical records. This hybrid slicing technique is rated against Privacy loss and Utility gain factors. Experimental results are analyzed to justify the performance of Slicing+ technique.