《计算机应用》唯一官方网站 ›› 0, Vol. ›› Issue (): 106-111.DOI: 10.11772/j.issn.1001-9081.2024020214

• 网络空间安全 • 上一篇    下一篇

融合用户偏好与差分隐私模型的位置隐私保护方法

朱亮1, 穆金巧1, 曹腾飞2(), 蔡增玉1, 张建伟1,3   

  1. 1.郑州轻工业大学 计算机科学与技术学院,郑州 450001
    2.青海大学 计算机技术与应用系,西宁 810016
    3.许昌职业技术学院 信息工程学院,河南 许昌 461000
  • 收稿日期:2024-03-05 修回日期:2024-04-17 接受日期:2024-04-24 发布日期:2025-01-24 出版日期:2024-12-31
  • 通讯作者: 曹腾飞
  • 作者简介:朱亮(1987—),男,河南焦作人,副教授,博士,CCF会员,主要研究方向:位置服务、隐私保护
    穆金巧(1998—),女,河南信阳人,硕士,主要研究方向:差分隐私、个性化隐私保护
    曹腾飞(1987—),男,湖北钟祥人,副教授,博士,CCF高级会员,主要研究方向:网络安全、隐私保护
    蔡增玉(1979—),男,河南鹤壁人,副教授,硕士,主要研究方向:网络安全、隐私保护
    张建伟(1971—),男,河南南阳人,教授,博士,主要研究方向:网络安全、内容分发。
  • 基金资助:
    国家自然科学基金资助项目(61902361);河南省重点研发专项(221111210500)

Location privacy preserving method by combining user preference and differential privacy model

Liang ZHU1, Jinqiao MU1, Tengfei CAO2(), Zengyu CAI1, Jianwei ZHANG1,3   

  1. 1.School of Computer Science and Technology,Zhengzhou University of Light Industry,Zhengzhou Henan 450001,China
    2.School of Computer Technology and Applications,Qinghai University,Xining Qinghai 810016,China
    3.School of Information Engineering,Xuchang Vocational Technical College,Xuchang Henan 461000,China
  • Received:2024-03-05 Revised:2024-04-17 Accepted:2024-04-24 Online:2025-01-24 Published:2024-12-31
  • Contact: Tengfei CAO

摘要:

位置社交网络(LBSN)将社交网络与地理位置相结合,为用户提供了新颖的个性化体验,而用户的位置隐私保护对LBSN系统的安全运行至关重要。针对位置隐私保护方法僵硬导致数据效用低、位置服务(LBS)体验质量下降的问题,提出一种融合用户偏好与差分隐私模型的位置隐私保护(UPDP-LPP)方法。首先,使用停留点提取算法获得用户的停留点集合;其次,使用特征融合方法标注停留点的类型;最后,在通过用户偏好来动态地获取隐私预算和噪声敏感度后,为隐私半径添加拉普拉斯噪声,从而保护用户的敏感位置信息。在两个公开的真实数据集上的实验结果表明,当隐私预算相同时,所提方法较TLDP (Trajectory Location Data Protection)、DPLPA (Differential Privacy-based Location Privacy protection Algorithm)和LPPM(Location Privacy Protection Mechanism)在隐私保护的数据效用上提高了10%以上。可见,UPDP-LPP不仅能保护用户位置隐私,而且提高了数据效用。

关键词: 位置社交网络, 位置服务, 用户偏好, 差分隐私, 位置隐私保护

Abstract:

Location-Based Social Network (LBSN) combines social network with geographical locations, providing users with novel personalized experiences. The protection of user location privacy is crucial for the secure operation of LBSN systems. To address the problem of rigid location privacy protection methods leading to low data utility and decreased quality of Location-Based Service (LBS) experiences, a User Preference-based and Differentially Private Location Privacy Protection (UPDP-LPP) method was proposed. Firstly, the set of user stay points was obtained by using a stay point extraction algorithm. Secondly, the types of stay points were labeled by using a feature fusion method. Finally, by dynamically obtaining privacy budget and noise sensitivity through user preferences, Laplace noise was added to the privacy radius to protect sensitive user location information. Experimental results on two public real datasets show that the proposed method improves the data utility of privacy protection by more than 10% compared to TLDP (Trajectory Location Data Protection), DPLPA (Differential Privacy-based Location Privacy protection Algorithm), and LPPM (Location Privacy Protection Mechanism) when the privacy budget is the same. It can be seen that UPDP-LPP not only protects user location privacy, but also enhances data utility.

Key words: Location-Based Social Network (LBSN), Location-Based Service (LBS), user preference, differential privacy, location privacy preserving

中图分类号: