Effective High Utility Itemsets Mining Algorithm for Incremental Database

Abstract

High-utility itemset mining (HUIM) majors have done a lot of research lately, the past few years. Almost all published algorithms focus on processing static databases, which do not utilize previously mined information to mine incremental databases. To solve this problem, some incremental HUIM algorithms were published and showed the possibility of development. In this study, a new algorithm named iHUIM based on the EIHI algorithm was improved. Unlike EIHI, which requires twice database scans, the iHUIM just scans the database only once. Additionally, using compact utility lists and some pruning strategies, iHUIM shows outperformance EIHI regarding the length of execution time and has a slight improvement in memory consumption.