Lossless and robust privacy preservation of association rules in data sanitization

Abstract

Data sanitization is a novel research area that conceals the sensitive rules given by the experts present in the original database with the appropriate modifications and then emancipates the modified database so that unauthorized persons cannot discover the sensitive rules and so the confidentiality of data is conserved against data mining methods. This paper primarily focuses on building an effective sanitizing algorithm for hiding the sensitive rules given by the experts/users. In order to minimize the four sanitization research challenges such as hiding failure, information loss, false rule generation and modification degree, the proposed method uses Firefly optimization algorithm. The proposed sanitization method has been compared and examined with other existing sanitizing algorithms depicting considerable improvement in terms of four research challenges that in turn can secure the selected database.