Student research abstract: Mining high average utility pattern using bio-inspired algorithm

Abstract

High average utility pattern (itemset) Mining (HAUIM) is a necessary research problem in the field of knowledge discovery and data mining. Several algorithms have been proposed to mine high average-utility itemsets (HAUIs). Nonetheless, the large search space leads to poor performance because of excessive execution time and memory usage. To handle this limitation, particle swarm optimization (PSO) is applied to mine HAUIs. In this paper, an effective Binary PSO-based algorithm namely HAUIM-BPSO is proposed to explore HAUI efficiently. In general, HAUIM-BPSO first sets the number of discovered potential high average-utility 1-itemsets (1-PHAUIs) as the size of a particle based on average utility upper bound (AUUB) property. The sigmoid function is also used in the updating process of the individual of the proposed HAUIM-BPSO algorithm. Substantial experiments conducted on publicly available datasets show that the proposed algorithm has better results than existing state-of-the-art algorithms in terms of runtime which can significantly reduce the combinational problem, memory usage, and convergence speed.