Least lion optimisation algorithm (LLOA) based secret key generation for privacy preserving association rule hiding. (1st July 2018)

• Record Type: Journal Article
• Authors: Menaga, D.; Revathi, S.
• Is Part Of: IET information security. Volume 12:Number 4(2018)
• Page Start: 332
• Page End: 340
• Publication Date: 2018-07-01
• Subjects: particle swarm optimisation -- data privacy -- data mining -- search problems -- public key cryptography; least lion optimisation algorithm -- secret key generation -- data mining -- PPDM techniques -- rule mining -- input database -- privacy factor -- particle swarm optimisation -- Cuckoo optimisation algorithm -- LLOA technique -- privacy preserving association rule hiding -- sensitive information hiding -- privacy preserving data mining -- fitness function -- sanitised database -- utility factor -- objective function -- search process -- firefly algorithm -- retail dataset; Computer security -- Periodicals; Cryptography -- Periodicals; Computer networks -- Security measures -- Periodicals; Database security -- Periodicals; 005.8
• DOI: 10.1049/iet-ifs.2017.0634
• Languages: English
• ISSNs: 1751-8709

Abstract

Abstract : Privacy preserving data mining (PPDM) is a well‐known research area, which can analyse the database ensuring privacy such that it does not expose the confidential details. This study proposes a technique of secret key generation for privacy preserving using the least lion optimisation algorithm (LLOA). The proposed algorithm involves two stages: rule mining and secret key generation for the sanitisation. Initially, whale optimisation algorithm mines the association rules for the input database and validates the rules with the newly formulated fitness function. An algorithm, LLOA is developed by modifying the lion optimisation algorithm (LOA) with the inclusion of least mean square (LMS) which generates a secret key to provide privacy in mining. With the secret key, LLOA converts the original database into the sanitised database. Then, the algorithm optimally selects a secret key such that the sanitised database hides sensitive information by the utilisation of two factors, namely, privacy factor and utility factor, in its objective function. From the results, it can be shown that the LLOA technique attains the privacy of 84.36, 83.74, and 82.46%; and utility of 81.37, 83.96, and 81.37%, for chess, T10I4D100K, and retail datasets.