Nowadays data mining is the way of extracting hidden knowledge from raw data whereas sequence mining aims to find sequential patterns that are frequent in the database, so publishing these data may lead to the disclosure of private information about organizations or individuals. Knowledge hiding is the process of hiding sensitive knowledge extracted previously from the database, to ensure that no abuse will be caused. This paper addresses the problem of sequential pattern hiding and proposes an efficient algorithm which uses a multi-objective approach to overcome the problem of sequence hiding as well as maintaining database fidelity as much as possible. It also shows that the proposed algorithm outperforms existing methods in terms of both speed and memory usage.

Keywords: data mining, sequence mining, knowledge hiding, sequential pattern

Full-Text [PDF 1458 kb]   (1353 Downloads)