An Enhanced Algorithm for Hiding Sensitive Association Rules Based on ISL and DSR Algorithms

Many privacy-preserving data mining algorithms attempt to hide what database owners consider as sensitive. Specifically, in the association rules domain, many of these algorithms are based on item restriction methods; that is, removing items from some transactions in order to hide sensitive frequent itemsets. There are two known algorithms for that purpose, ISL (Increase Support of Left) and DSR (Decrease Support of Right). Both of them make use of user-specified values for minimum support threshold and minimum confidence threshold as input. Since ISL and DSR techniques aim at hiding all sensitive rules, they cannot avoid the undesired side effects. In this paper, a new algorithm for hiding sensitive rules is proposed based on ISL and DSR. It depends on decreasing the confidence of the sensitive rule by dealing with both Left Hand Side (LHS) and Right Hand Side (RHS) according to the ratio between them. This technique introduces introduce a reasonable side effect (lost and new rules) as compared with ISL and DSR algorithms which introduces a high number of new rules and/or lost rules. Experimental results have shown that the proposed algorithm effectively reduces the side effects which occur due to the hide operation when compared with the ISL and DSR algorithms and that it gives a good hiding ratio for the sensitive rules