Privacy preserving high utility mining based on genetic algorithms

In this paper, a GA-based privacy-preserving utility mining method is proposed to delete appropriate transactions for hiding sensitive high utility itemsets from a database. The downward closure property and the pre-large concepts are adopted in the proposed algorithm to reduce the cost of rescanning databases. Experiments are also conducted to evaluate the performance of the proposed approach in execution time and the amount of side-effects. © 2013 IEEE.