Position big data differential privacy division publishing method based on non-uniform quadtree

Abstract

The invention discloses a position big data differential privacy division publishing method based on a non-uniform quadtree, mainly solving the defect of poor comprehensive performance of the existingposition information publishing method, and improving the publishing privacy protection strength and the range counting query precision of position big data. The method comprises the following steps:firstly, determining a division structure and a division depth by analyzing distribution characteristics of a to-be-published position big data set, and carrying out non-uniform quadtree iterative division on a two-dimensional region which does not meet a uniformity condition according to a depth-first principle until a stop condition of an algorithm is met; then, distributing differential privacy budget to each sub-region in the division structure according to the principle of firstly longitudinally geometrically distributing and then transversely adjusting the proportion, firstly, calculating geometric privacy budget values needing to be distributed to all layers according to the division depth of a non-uniform quadtree, and then conducting local adjustment of privacy budget according to the density proportion of four nodes in the same subtree; and finally, adding the original statistical value in each divided region to the differential privacy noise to obtain final published data.

Description

technical field [0001] The invention belongs to the field of big data application and privacy protection, and specifically relates to a location big data differential privacy partitioning publishing method based on a non-uniform quadtree structure, which can be used in big data publishing systems, intelligent transportation systems, Internet of Vehicles, LBS systems, Centralized location big data release and privacy protection in many application fields such as location advertisement push. Background technique [0002] As the "natural entrance" of the Internet, location big data is widely used in mobile Internet, Internet of Vehicles, intelligent transportation systems, location-based service systems, social networks and other popular areas of big data. The user's location information is aggregated, fused, analyzed and processed by the centralized big data publishing platform, and provided to scientific research, decision support and public services in various industries acc...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to address the deficiencies of the above-mentioned existing methods, and propose a location big data differential privacy partition publishing method based on a non-uniform quadtree structure, so as to improve the defect of poor comprehensive performance of the existing partition publishing method and improve location big data. Published Privacy Preservation Strength and Range Count Query Precision of

Images