基于差分隐私的混合位置隐私保护

针对现有差分隐私k-means算法对初始中心点敏感、用户位置数据误差偏大、可用性较低等问题,根据LBS的特点,引入人流密度的概念,提出一种基于差分隐私k-means的混合位置隐私保护方法。根据LBS特点将用户位置点分成离散位置点和非离散位置点,基于差分隐私技术,采用改进聚类算法对位置信息进行泛化和加噪;通过分析用户位置点的稀疏程度来确定离散点,对离散点位置信息采用基于差分隐私的单独加噪技术;对非离散点采用基于差分隐私的改进k-means算法进行泛化处理,以实现用户位置信息的隐私保护。仿真实验表明,在相同隐私预算的前提下,该方法具有较高的数据可用性。

HYBRID LOCATION PRIVACY PROTECTION BASED ON DIFFERENTIAL PRIVACY

Aiming at problems such as the existing differential privacy k-means algorithm is sensitive to the initial center point,the user location data error is large,and the availability is low,etc.,according to the characteristics of LBS,the concept of crowd density was introduced,and a hybrid location privacy protection method based on differential privacy k-means was proposed.In terms of the characteristics of the LBS,the location of the user s location was divided into discrete points and non-discrete points.Base on the differential privacy technology,using improved clustering algorithms to generalize and enhance the location information,the discrete points were determined by analyzing the sparsity degree of the user s position points.The improved k-means algorithm based on differential privacy was used to generalize the non-discrete points to realize the privacy protection of user location information.Simulation results show that the proposed method has high data availability under the premise of the same privacy budget.