%A HUANG Siyong, CHEN Tingting, LU Qing, WU Yingjie, YE Shaozhen %T Differentially privacy two-dimensional dataset partitioning publication algorithm based on kd-tree %0 Journal Article %D 2015 %J Journal of Shandong University(Engineering Science) %R 10.6040/j.issn.1672-3961.2.2014.120 %P 24-29 %V 45 %N 1 %U {http://gxbwk.njournal.sdu.edu.cn/CN/abstract/article_1117.shtml} %8 %X The existing differentially privacy two-dimensional dataset partitioning publication approaches might result in over-partitioning of certain local regions thus lower the query accuracy. To solve this problem, a new differentially privacy partitioning publication algorithm based on kd-tree for two-dimensional dataset (kd-PPDP) was proposed.To reduce the noise generated by the Laplace mechanism and improve the accuracy of query, the new approach which was inspired by the core thought of kd-tree took the gridding data distribution into account and merged adjacent grid units with similar information heuristically to prevent local over-partitioning. The new approach was compared with the existing differentially privacy partitioning publication approachs based on grid structure in terms of query error and time complexity. Experimental results showed that the approach was feasible and effective.