Personalized Privacy-Preserving with high performance: (α, ε)-anonymity

Express data contains valuable information that can be used for optimizing goods transporting and delivering, but it also has the risk of individual privacy leakage during data publishing and analyzing especially after implementing realname registration policy in China. Due to some special traits of express data, traditional methods have some drawbacks that failing to deal with it. So it is significant to find a personalized privacy-preservation in regard to express data. In this paper, we innovatively propose a ($\alpha \varepsilon$)-anonymity model for protecting privacy in express data which is based on Lossy-join. After anonymizing, we project the data into two tables, which not only meets the constraint of ($\alpha \varepsilon$)-anonymity, but also consider the characteristics of express data, meanwhile, we handle this model with high performance and data utility. At last, we conduct experiments on real-world express data to verify the ability of our approach and algorithm.