Privacy Preserving in Association Rule Mining

Abstract

Association rule mining is one of the most important techniques of data mining that are used to extract the association patterns from large databases. Association rules are one of the most important assets of any organization that can be used for business development and profitability increase. Association rules contain sensitive information that threatens the privacy of its publication and they should be hidden before publishing the database. The aim of hiding association rules is to delete sensitive association rules from the published database so that possible side effects are reduced. In this paper, we present a heuristic algorithm DCR to hide sensitive association rules. In the proposed algorithm, two clustering operations are performed on the sensitive association rules and finally, a bunch of smaller clusters is chosen to hide. A selection of a smaller bunch of clusters reduces the changes in the database and side effects. The results of performing experiments on real databases, shows the impact of the proposed algorithm on missing rules reduction.