Protecting Privacy by Multi-dimensional K-anonymity

Abstract

Privacy protection for incremental data has a great effect on data availability and practicality. K-anonymity is an important approach to protect data privacy in data publishing scenario. However, it is a NP-hard problem for optimal k-anonymity on dataset with multiple attributes. Most partitions in k-anonymity at present are single-dimensional. Now research on k-anonymity mainly focuses on getting high quality anonymity while reducing the time complexity, and new method of realization of k-anonymity properties according to the requirement of published data. Although most k-anonymity algorithms perform well on static data, their effects decrease when they are on the changing data of real world. This paper proposes a multi-dimensional k-anonymity algorithm based on mapping and divide-and-conquer strategy that is feasible and performs much better in k-anonymity. The second main contribution of this paper is an effective k-anonymity method based on incremental local update on large dataset. It incrementally updates the changing dataset, and a threshold is set to assure the stability of update. Neighbor equivalence sets and similar equivalence sets are computed by their position, which not only avoids the cost of recalculation aroused by little change of dataset, but also improves the practical application performance since the dataset satisfies k-anonymity properties. The experiment shows that the proposed algorithm has a better performance in both time cost and anonymity quality, compared to the methods at present.