CISA and FISA: Efficient algorithms for hiding association rules based on Consequent and Full Item Sensitivities

Database sharing is a necessary task for most of the real world organizations in order to take advantages from organizational cooperation or to find out hidden knowledge buried in their databases, so they can make decisions more precisely. This incurs the risk of sensitive knowledge disclosure which may occur due to malicious use of database. In this paper, we propose two heuristic algorithms called CISA and FISA (Consequent and Full Item Sensitivities Algorithms) for hiding sensitive association rules based on new terms called Consequent and Full Item Sensitivities. This algorithm can hide multiple sensitive rules at the same time. To illustrate efficiency of our work, we compare our algorithms with Algo 1.a which is a popular algorithm for hiding association rules.