HIDING FREQUENT ITEMSETS BASED ON INTEGER LINEAR PROGRAMMING METHOD COMBINED WITH IDEAL POSITIVE BORDER

Abstract

This study proposes a method to hide sensitive frequent itemsets in transaction databases. This proposed method is based on using information from the ideal positive border to build integer linear programming equations. The solution of this equation determines the transactions that need to be sanitized to completely hide the sensitive frequent itemsets. In case the equation has no solution, some coefficients are added to loosen the constraints of the problem. Experimental evaluation of this method on some well-known data sets shows that this method has higher accuracy than the method using traditional integer linear programming.