A Privacy-preserving Location Recommendation Scheme without Trustworthy Entity
Changli Zhou, Jing Peng, Ying Ma, Qingfeng Jiang · 2021 IEEE 20th International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom) · 2021
The widespread application of location recommendation service brings convenience and social fun to users, but the misuse of user check-in data and location recommendation results can bring serious privacy leakage, because these check-in data and location recommendation results imply private information of users' interests and behavioral hobbies. To address the check-in privacy and recommendation result privacy leakage issues in existing location recommendation schemes, we first propose a location recommendation algorithm based on the similarity of user historical check-ins, and then we devise three location recommendation protocols on ciphertext to protect the privacy of user check-in data and recommendation results based on the homomorphic property of Paillier cryptosystem without assuming any trustworthy entity. These protocols calculate the ciphertexts of recommendation results at the location recommendation server(LSP) without learning any explicit check-in privacy of the user, and it also guarantees that the decryption server(DS) can only implement the decryption operations as requested by the user without being aware of any decrypted recommendation results, so that the privacy of user check-in data and recommendation results are both preserved solidly in our protocols. Although the above recommendation process is carried out without any user's explicit data, it is still effective enough to guarantee the accuracy of recommendation results. Finally, we prove the security, efficiency and accuracy of our scheme by extensive performance analysis.