Abstract:

Abstract High Utility Itemset Mining (HUIM) is the progression of identifying highly profitable items by considering the unit profit of the item from the huge transactional database. HUIM is an essential subject with broad applications in recent years. HUIM paves the way to know the profitable items using factors namely profit and quantity. Until today, abundant algorithms have been found to mine High Utility Itemset (HUI) and it is entirely different from the conventional mining algorithms. Most of the utility mining algorithms generate the itemset recurrently and scan the database redundantly, which leads to computational complexity. To overcome this issue, a pruning technique is introduced with a matrix and Frequent Pattern (FP) tree is constructed with the pruned matrix whereby the complexity in HUI identification is minimized. Experimental results are investigated using a benchmark dataset and the outcome depicts that the proposed pruned matrix-aided utility tree (PMAUT) outperforms the existing state of art techniques in terms of time consumption and memory usage.

Keywords: