Fully privacy-preserving location recommendation in outsourced environments

Abstract

Currently, location-based services (LBS) have been widely used in real-world settings, including restaurant and travel recommendations. To reduce workload and improve query efficiency, a service provider usually outsources its services to a powerful cloud server. However, the service provider’s database and users’ queries always contain sensitive information, so their leakage to the cloud may raise serious privacy concerns. Although some existing schemes have been proposed to address the privacy problems, they are impractical in real-world LBS due to some issues in privacy, accuracy, or heavy computation costs for query users. In order to overcome these problems, we propose a fully privacy-preserving location recommendation scheme that supports multi-attribute queries and returns accurate results based on the recommendation condition. Specifically, based on the Paillier cryptosystem, we first propose a secure equal test (SET) protocol to check whether two encrypted values are equal. Second, with our proposed protocols, we develop a privacy-preserving location recommendation scheme without revealing anything about the service provider or query users. Finally, we analyze the security of our scheme in the semi-honest model and show that the privacy of the service provider and query users is well protected. Meanwhile, we evaluate the performance of our scheme using synthetic datasets. The experimental results demonstrate that our proposed scheme is practical in real-world applications.