A differential privacy histogram publishing method and system that prioritizes the availability of small buckets

Abstract

The present invention provides a differential privacy histogram publishing method and system that prioritizes the availability of small buckets, and performs preliminary perturbation, including preliminarily perturbing the original histogram with a part of the privacy budget combined with the Laplace differential privacy mechanism, and obtaining the preliminary perturbation. Intermediate histogram; threshold function processing, including thresholding the intermediate histogram after initial disturbance; sorting, including sorting the updated intermediate histogram to obtain a histogram sorted from small to large; grouping and clustering, including using residual The privacy budget of , grouping the histograms sorted from small to large to obtain a grouping set; publishing, including generating the final noise for each group in the grouping set in combination with the remaining privacy budget, and obtaining the perturbed histogram after perturbation image and used for publication. The present invention not only satisfies strict ε-differential privacy protection, but also achieves a balance between reconstruction error and noise error.

Description

technical field [0001] The invention belongs to the field of information security, and in particular relates to a differential privacy histogram publishing method and system that give priority to the availability of small buckets. Background technique [0002] With the rapid development of the mobile Internet and the widespread popularity of mobile devices, a large amount of data is generated every day based on various App applications. Although the collection, extraction and publication of data can help users obtain the required information efficiently, quickly and accurately from the complex data, the privacy and security issues are also becoming more and more serious. The potential or direct harm caused by the leakage of these private information to users makes the privacy and security of users' personal become a hot topic. [0003] The issue of data privacy protection was first raised by the statistician Dalenius in the late 1970s. He believes that to protect the priva...

AI Technical Summary

Problems solved by technology

[0013] The problem to be solved by the present invention is to provide a relatively balanced unit noise amount for both small buckets and large buckets under the background of uneven data distribution for traditional histogram publishing that satisfies ε-differential privacy

Images