Verifiable privacy-preserving association rule mining using distributed decryption mechanism on the cloud. (1st September 2022)

• Record Type: Journal Article
• Authors: Chen, Yange; Zhao, Qingqing; Duan, Pu; Zhang, Benyu; Hong, Zhiyong; Wang, Baocang
• Is Part Of: Expert systems with applications. Volume 201(2022)
• Publication Date: 2022-09-01
• Subjects: Verifiable -- Association rule mining -- Privacy-preserving -- Active attack -- Poisoning attack; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2022.117086
• Languages: English
• ISSNs: 0957-4174

Abstract

Abstract: As one of the important ways in data mining, the association rule mining is to analyze the correlation of transactions based on massive data and mine the hidden valuable information. However, excessive data collection and analysis might lead to the privacy leakage of user data and the damage of data integrity. Meanwhile, in the existing privacy-preserving schemes, the cloud servers such as the evaluator have strong decryption capabilities, which causes active attacks easily, or data reliability in mining and analysis is not considered. In response to these problems, this paper proposes a verifiable privacy-preserving association rule mining scheme (VPPARM) using distributed decryption mechanism on the cloud. First, the scheme adopts distributed decryption to complete the data mining tasks through the dual-cloud servers, which weakens the decryption ability of the cloud servers and prevents the server from active attacks. Secondly, to protect the privacy of the association rule mining process, our scheme adopts adding virtual transactions, permutations, and random number masking collaboratively to hide the data in the whole mining process. In addition, to eliminate the hidden danger of illegal users, our scheme adopts a verifiable short digital signature scheme to verify data integrity and ensure data reliability, avoiding the poisoning attack. Finally, through the performance evaluation, the results demonstrate that our scheme realizes correctness, security, reliability with lower communication and computation costs and improves efficiency to a certain extent. Highlights: This paper proposes a verifiable privacy-preserving ARM scheme. This scheme stores two weak private keys realizing jointly decryption ciphertext. The BLS signature scheme verifies data integrity and legitimacy. The CSP uses adding virtual transactions, permutation, and random number methods.