面向稠密区域的本地化差分隐私自适应空间分解

打开文本图片集

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 中图分类号：TP309.2 文献标志码：A 文章编号：1001-3695（2025）08-036-2518-07 doi：10.19734/j.issn.1001-3695.2024.11.0505

Adaptive spatial decomposition based on localized differential privacy for dense regions

Ji Bo，Li Xiaohui†，JiaXu （SchoolofElectronics&Information Enginering，Liaoning UniversityofTechnology，JinzhouLiaoning21oo，China）

Abstract：Toaddress theisues oflowqueryaccuracyandeficiencyin processng spatialdatausing conventionaluniform gridmethodsandadaptive griddecompositionmethods，，this paperdevelopedanadaptive spatialdecompositionalgorithmbased onlocal diffrential privacy（LDP-ASDT）.LDP-ASDTperformed spatial decomposition hierarchicallyusing agrouping strategyto separate denseandsparseregions.For dense regions，it further adaptively decomposed them bysetting appropriate thresholds using aquadtre.Itperturbedthe decompositionresults using one-dimensionalcoding toachieve privacy protection. Theoreticalanalysisdemonstratesthatthisalgorithmsatisfieslocalizeddiferentialprivacy.Experimentsconductedonthree real datasetsshowthatthequeryacuracyandoperationaleficiencyofthisalgorithmaresuperiortothoseofGT-R，PrivAG， KDRQ，and ASDQT algorithms.Specificall，compared to the ASDQT algorithm，LDP-ASDT algorithm doubles thequery accuracy and increases the operational rate by 17% in dense regions.

Key words：localdiferential privacy；adaptivespatialdecomposition；adaptivegriddecomposition；randomizedresponse; spatial range query

0 引言

大数据时代，随着信息技术的迅猛发展和智能设备的广泛普及，空间数据因其巨大的潜在价值而备受关注。（剩余16041字）