Advanced cluster-based attribute slicing: A new approach for privacy preservation

Abstract

Privacy preservation data is an emerging field of research in the data security. Numerous anonymization approaches have been proposed for privacy preservation such as generalization and bucketization. Recent works show that both the techniques are not suitable for high-dimensional data publishing. Several other challenges for data publishing are speed and computational complexity. Slicing is a novel anonymization technique which partitions the data both horizontally and vertically and solves the problem of high-dimensional complexity. To overcome the other challenges of speed and computational complexity, a new advanced clustering algorithm is used with slicing. This advanced clustering technique is used to partition the attributes into vertical columns and tuple grouping algorithm for horizontal partitioning. The experimental results confirm that the advanced clustering algorithm improves the speed of clustering accuracy, reduces the computational complexity, and the outcome of this work is resilient to membership, identity, and attributes disclosure.