A MaxMin approach for hiding frequent itemsets

Abstract

In this paper, we are proposing a new algorithmic approach for sanitizing raw data from sensitive knowledge in the context of mining of association rules. The new approach (a) relies on the maxmin criterion which is a method in decision theory for maximizing the minimum gain and (b) builds upon the border theory of frequent itemsets. Experimental results indicate the effectiveness of the proposed methodology both with respect to the hiding results as well as with respect to the time performance compared to similar state of the art approaches.