An Efficient Candidate Pruning Technique for High Utility Pattern Mining

Abstract

High utility pattern mining extracts more useful and realistic knowledge from transaction databases compared to the traditional frequent pattern mining by considering the non-binary frequency values of items in transactions and different profit values for every item. However, the existing high utility pattern mining algorithms suffer from the level-wise candidate generation-and-test problem and need several database scans to mine the actual high utility patterns. In this paper, we propose a novel tree-based candidate pruning technique HUC-Prune (high utility candidates prune) to efficiently mine high utility patterns without level-wise candidate generation-and-test. It exploits a pattern growth mining approach and needs maximum three database scans in contrast to several database scans of the existing algorithms. Extensive experimental results show that our technique is very efficient for high utility pattern mining and it outperforms the existing algorithms.