CCMine: Efficient Mining of Confidence-Closed Correlated Patterns

Correlated pattern mining has become increasingly important recently as an alternative or an augmentation of association rule mining. Though correlated pattern mining discloses the correlation relationships among data objects and reduces significantly the number of patterns produced by the association mining, it still generates quite a large number of patterns. In this paper, we propose closed correlated pattern mining to reduce the number of the correlated patterns produced without information loss. We first propose a new notion of the confidence-closed correlated patterns, and then present an efficient algorithm, called CCMine, for mining those patterns. Our performance study shows that confidence-closed pattern mining reduces the number of patterns by at least an order of magnitude. It also shows that CCMine outperforms a simple method making use of the the traditional closed pattern miner. We conclude that confidence-closed pattern mining is a valuable approach to condensing correlated patterns.