基于敏感度判定的位置隐私保护方法  

Location Privacy Protection Method Based on Sensitivity Determination

作　　者：刘琨[1] 王希孔 王辉[1] 周超 刘沛骞[1] LIU Kun;WANG Xi-kong;WANG Hui;ZHOU Chao;LIU Pei-qian(School of Computer Science and Technology,Henan Polytechnic University,Jiaozuo 454002,China)

机构地区：[1]河南理工大学计算机科学与技术学院,河南焦作454002

出　　处：《小型微型计算机系统》2023年第11期2450-2456,共7页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61300216)资助。

摘　　要：在位置隐私保护中,差分隐私保护机制通过随机函数对真实位置加噪,从而保护真实位置信息.然而在隐私预算分配问题上往往易造成资源浪费以及隐私保护效率低下.针对此问题,本文在已有研究的基础之上提出一种基于概率相似性度量的差分隐私算法DPBO.首先,使用加权有向图来收集获取历史位置信息;然后通过范围度量找出历史位置点与真实轨迹中每个位置点R所对应的相似位置点集SET,对SET与R进行一对多(One To Many, OTM)概率相似性度量,得出每个R点的位置敏感度ΔG;最后根据不同ΔG,为真实轨迹中R分配相应隐私预算,添加Laplace噪声.通过实验,证明了该方案具有数据可用性和可行性.In location privacy protection,the differential privacy protection mechanism protects the real location information by adding noise to the real location with a random function.However,the privacy budget allocation problem often leads to waste of resources and inefficiency of privacy protection.To solve this problem,on the basis of the existing research,this paper proposes a differential privacy algorithm DPBO based on probabilistic similarity metric.First,we use a weighted directed graph to collect and obtain the historical location information;then,we use the range metric to find the similar location point set SET corresponding to each location point R in the historical location and the real trajectory,and we perform the One-To-Many(OTM)probabilistic similarity metric on SET and R,Derive the position sensitivity of each R point AG;finally,according to the different AG,the corresponding privacy budget is assigned to R in the real trajectory and Laplace noise is added.Through experiments,the scheme is demonstrated to have data availability and feasibility.

关 键 词：差分隐私 加权有向图 相似位置点集 位置敏感度 位置隐私保护 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]