轻量级位置感知推荐系统隐私保护框架

摘要

As one of the most important method to provide individual location-based service for users,location-aware recommender system has recently experienced a rapid development and draw significant attention from the research community.Such problems are more prominent when service providers,who have limited computational and storage resources,leverage on cloud platforms to fit in with the tremendous number of service requirements and users.However,state-of-the-art work either suffers from an inaccurate recommendation quality or low efficiency.To address the problem,a lightweight privacy-preserving framework for location-aware recommender system is proposed.Before requesting the service,service providers should firstly outsource the historical evaluation information,which is processed with random function,to cloud platform through the framework.Then,the similarities of information are computed securely with the help of cloud platform.When requesting the service,recommendation users encrypt the interesting area with comparable encryption and send the encrypted results to service providers.After that,the service providers will predict the recommendation results with the help of cloud platform through a carefully designed secure protocol.We have also theoretically proved that user information is private and will not be leaked during a recommendation.Finally,empirical results over a real-word dataset demonstrate that our framework can efficiently recommend POIs with a high degree of accuracy in a privacy-preserving manner.Compared with the existing work based on homomorphic encryption methods,our lightweight scheme is also more efficient and has a better user experience.%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.