An Approach to Decrease Execution Time and Difference for Hiding High Utility Sequential Patterns

Abstract

Nowadays, databases are shared commonly in various types between companies and organizations. The essential requirement is to release aggregate information about the data, without leaking individual information about participants. So the Privacy Preserving Data Mining (PPDM) has become an important research topic in recent years. PPDM models are applied commonly on hiding association rule, hiding high utility itemsets mining and also on hiding High Utility Sequential Patterns (HUSPs) mining. The goal of hiding utility sequential patterns is to find the way to hide all HUSPs so that the adversaries cannot mine them from the sanitized database. The exiting researches hasn’t considered in details about the difference ratio between the original database and the sanitized database after hiding all HUSPs. To address this issue, this paper presents two algorithms, which are HHUSP-D and HHUSP-A (Hiding High Utility Sequential Pattern by Descending and Ascending order of utility) to decrease the difference and also decrease execution time. In the proposed algorithms, a additional step is added to the exiting algorithm, HHUSP, to rearrange the hiding order of the HUSPs. Experimental results show that HHUSP-D is better performance than HHUSP [4] not only on the difference but also on the execution time.