轻量级位置感知推荐系统隐私保护框架

发布时间：2018-09-17 13:10

【摘要】：作为提供个性化位置服务的一种重要手段,高速、高效的位置感知推荐服务成为当前研究的热点.涉及多方参与的传统推荐流程存在着用户私密信息复制、盗取等安全威胁,给用户的隐私保护带来了新的挑战,尤其是当服务提供者将数据外包给第三方云平台时,隐私泄露问题会更加凸显.然而,现有的解决方案均存在推荐质量低、响应速度慢的问题.为解决上述问题,提出了一种轻量级的位置感知推荐系统隐私保护框架.利用该框架,服务提供者将随机处理后的历史评价信息外包给云平台,并通过安全协议在云平台的辅助下进行相似度信息的安全计算;同时,推荐用户利用可比较加密将其感兴趣的位置区域进行加密并发送给云平台进行请求服务,并通过安全协议实现推荐结果的安全预测.最后,通过在真实数据集中进行仿真调试,结果表明该框架能够在保证用户隐私安全的前提下,准确、高效地为用户推荐位置点.同时,与同态加密方案相比,该方案更加高效,能够更快速地响应用户的请求.
[Abstract]:As an important means to provide personalized location services, high speed and efficient location awareness recommendation services have become the focus of current research. In the traditional recommendation process involving multi-participation, there are security threats such as private information replication and theft, which brings new challenges to the privacy protection of users, especially when service providers outsource data to third party cloud platforms. The issue of privacy leaks will become more prominent. However, the existing solutions have the problems of low recommendation quality and slow response speed. In order to solve the above problems, a lightweight privacy protection framework for location-aware recommendation systems is proposed. Using this framework, service providers outsource randomly processed historical evaluation information to the cloud platform, and calculate the security of similarity information through security protocols assisted by the cloud platform. The recommended users use comparable encryption to encrypt and send to the cloud platform to request service, and realize the security prediction of the recommended results through the security protocol. Finally, through simulation and debugging in real data sets, the results show that the framework can accurately and efficiently recommend location points to users on the premise of ensuring the privacy of users. At the same time, compared with homomorphic encryption scheme, this scheme is more efficient and can respond to user's request more quickly.
【作者单位】：西安电子科技大学计算机学院;西安电子科技大学网络与信息安全学院;中央财经大学信息学院;
【基金】：国家自然科学基金(61202179,U1405255,61502368,61602537,61602357,61672413,U1509214,U1135002) 国家“八六三”高技术研究发展计划项目基金(2015AA016007) 陕西省自然科学基金(2015JQ6227,2016JM6005) 国家111计划(B16037) 中央高校基本科研业务费(JB150308,JB150309,JB161501)资助~~
【分类号】：TP309;TP391.3

【相似文献】