Adaptive Grid Decomposition Algorithm based on Standard Deviation Circle Radius

Abstract

The differential privacy preservation model based on spatial dataset meshing has been widely concerned, but the distribution characteristics of the dataset and user's query granularity are often ignored or not fully considered in the partitioning of the dataset. Aiming at deficiencies in existing mesh-based algorithms, a standard deviation circle radius adaptive grid decomposition (SDCAG) algorithm is proposed. Firstly, the standard deviation circle radius is introduced to quantitatively represent the distribution characteristics of datasets in order to calculate privacy preservation requirements. Secondly, filtering and bucketing are used to reduce the noise error. Finally, the improved query precision is implemented based on the post-processing. Experiments on the NYC dataset, the Beijing dataset, and the Checkin dataset show that the SDCAG algorithm is superior to similar algorithms in terms of query performance.