Keyword-Based PIR Method for Enhancing Large-Scale Power Credit

Power credit reports are an important means of enhancing enterprises’ credit. However, there are two main problems in existing inquiry services. The first problem is that existing methods do not protect the inquiry preference, which is also private information. The second problem is that existing private information retrieval methods cannot be extended to larger-scale databases. To address these issues, a novel inquiry method named “Effi-Retrieval” is proposed to simultaneously ensure the security and efficiency of inquiries. Specifically, Paillier homomorphic encryption and oblivious polynomial calculations are used to ensure the security of inquiries. An optimal binning strategy, along with the k-anonymity method and hash mapping, is employed to meet the personalized privacy requirements of the querying party in practice. Finally, a security analysis of Effi-Retrieval is presented, and numerical experiments are conducted to demonstrate the impacts of user privacy requirements and an affordable fee on communication overhead. The ablation experimental results showed that, in comparison to settings without the binning strategy or k-anonymity method, Effi-Retrieval can significantly reduce retrieval time.