A differential privacy multidimensional data release model

As a new privacy notion, differential privacy, a hot topic in the privacy preserving of data release and mining, has grown in popularity recently. In order to protect the privacy of users, we need to find a differential privacy multidimensional data release schema. The paper gives a simple differential privacy multidimensional data release model based on the improved kd-Tree algorithm. Firstly, divide the data cube into several parts using the kd-Tree algorithm. In this step, the information entropy was introduced to select the partition dimension to improve the utility of kd-Tree algorithm. Secondly, add the noise information to the partitioned database with the Laplace mechanism. And then, it is provided that the model could ensure differential privacy while retaining the utility of the common queries.