Protecting Privacy While Discovering and Maintaining Association Rules

The k-anonymity is an efficient model to preserve data privacy. Of late, this model has been applied to the area of privacy-preserving data mining but the state-of-the-arts are still far from practical needs. In this paper, we propose a new approach that preserves privacy and maintains data utility in data mining. Concretely, we use a k-anonymity model to preserve privacy while discovering and maintaining association rules through a novel algorithm, M3AR-member migration technique for maintaining association rules. We do not use the existing generalization and suppression techniques to achieve a k-anonymity model. Instead, we propose a member migration technique that is more appropriate for the requirements of maintaining association rules. Experimental results establish the practical value and theoretical analyses of our new technique.