A heuristic algorithm for quick hiding of association rules

Abstract

Increasing use of data mining process and extracting of association rules caused the introduction of privacy preserving in data mining. A complete publication of the database is inconsistent with security policies and it would result in disclosure of some sensitive data after performing data mining. Individuals and organizations should secure the database before the publication, because if they neglect this issue they will be harmed. The owners of database consider factors such as database size, precision in immunization and velocity in choosing the right approach in order to hide the association rules. Besides the large volume of data and precision in immunization, we should optimize the time of operation and this is one of the issues that has received a little attention. In this paper, FHA algorithm is introduced for hiding sensitive patterns. In this algorithm, it is being tried to reduce the overload of ordering transactions by decreasing database scans. Also, we have reduced the side effects by selecting the appropriate item for performing the modifications. Conducted experiments indicate the execution of this algorithm in appropriate hiding of sensitive association rules.