摘要
针对常规均匀网格法和自适应网格分解法处理空间数据时存在查询精度与查询效率较低的问题,提出一种基于本地化差分隐私的自适应空间分解算法(LDP-ASDT)。通过分组策略对空间进行分层分解,划分出稠密与稀疏区域;利用四分树通过设定合适阈值,对稠密区域进一步自适应分解;利用一元编码对划分结果进行扰动实现隐私保护。理论分析证明该算法满足本地化差分隐私。在三个真实数据集上进行实验,结果表明,查询精度与运行效率优于GT-R、PrivAG、KDRQ、ASDQT算法,在稠密区域查询精度比ASDQT算法提升一倍,运行速率提升17%。
To address the issues of low query accuracy and efficiency in processing spatial data using conventional uniform grid methods and adaptive grid decomposition methods,this paper developed an adaptive spatial decomposition algorithm based on local differential privacy(LDP-ASDT).LDP-ASDT performed spatial decomposition hierarchically using a grouping strategy to separate dense and sparse regions.For dense regions,it further adaptively decomposed them by setting appropriate thresholds using a quadtree.It perturbed the decomposition results using one-dimensional coding to achieve privacy protection.Theoretical analysis demonstrates that this algorithm satisfies localized differential privacy.Experiments conducted on three real datasets show that the query accuracy and operational efficiency of this algorithm are superior to those of GT-R,PrivAG,KDRQ,and ASDQT algorithms.Specifically,compared to the ASDQT algorithm,LDP-ASDT algorithm doubles the query accuracy and increases the operational rate by 17%in dense regions.
作者
季博
李晓会
贾旭
Ji Bo;Li Xiaohui;Jia Xu(School of Electronics&Information Engineering,Liaoning University of Technology,Jinzhou Liaoning 121000,China)
出处
《计算机应用研究》
北大核心
2025年第8期2518-2524,共7页
Application Research of Computers
基金
国家自然科学基金青年基金资助项目(61802161)
辽宁省应用基础研究计划资助项目(2022JH2/101300278,2022JH2/101300279)
2024年辽宁省属本科高校基本科研业务费专项资金资助项目(LJZZ212410154025,LJZZ222410154004)
辽宁工业大学研究生教育改革创新项目(YJG2023013)。
关键词
本地化差分隐私
自适应空间分解
自适应网格划分
随机响应
空间范围查询
local differential privacy
adaptive spatial decomposition
adaptive grid decomposition
randomized response
spatial range query
