Research of Data Publishing Based on Micro-aggregation

Abstract

In order to solve the security problem and huge loss of data that exists in traditional K-Anonymity model. Basing on P-sensitive-k Anonymous Model, this article has proposed a new data publishing algorithm which is on the basis of cluster. This algorithm firstly divides the sensitive attribute into different groups by sensitiveness and take different constraint on the frequency that the same sensitive attribute group occurs in the same equivalence class, and then group the data table according to Micro-aggregation thought, thus solve the security problems happens in traditional anonymous models like homogeneity attack, background knowledge attack, possibility attack and so forth.