More Efficient Algorithm for Mining Frequent Patterns with Multiple Minimum Supports

Abstract

Frequent pattern mining (FPM) is an important data mining task, having numerous applications. However, an important limitation of traditional FPM algorithms, is that they rely on a single minimum support threshold to identify frequent patterns (FPs). As a solution, several algorithms have been proposed to mine FPs using multiple minimum supports. Nevertheless, a crucial problem is that these algorithms generally consume a large amount of memory and have long execution times. In this paper, we address this issue by introducing a novel algorithm named efficient discovery of Frequent Patterns with Multiple minimum supports from the Enumeration-tree (FP-ME). The proposed algorithm discovers FPs using a novel Set-Enumeration-tree structure with Multiple minimum supports (ME-tree), and employs a novel sorted downward closure (SDC) property of FPs with multiple minimum supports. The proposed algorithm directly discovers FPs from the ME-tree without generating candidates. Furthermore, an improved algorithms, named \({\text {FP-ME}}_\mathrm{DiffSet}\), is also proposed based on the DiffSet concept, to further increase mining performance. Substantial experiments on real-life datasets show that the proposed approaches not only avoid the “rare item problem”, but also efficiently and effectively discover the complete set of FPs in transactional databases.