Mining Condensed Non-Redundant Level-Crossing Approximate Association Rules

In association rule mining one intractable problem is the huge number of the extracted rules, especially, in the case of level-crossing association rules. In this paper, aiming at the redundancy produced during level-crossing association rules mining, an approach for eliminating level-crossing approximate redundant rules is proposed. In the method, the redundancies are divided combination with the dataset’s hierarchy or taxonomy into two categories: hierarchical Self-Redundancy and Inter-Redundancy, thus in the mining processing, deleting the Self-Redundant rules, removing the redundant rules from the InterRedundancy based on their definitions and characters in respective steps. The experiments show that the number of the extracted rules has been considerably reduced. Keywords-component; level-crossing association rules; redundant rules; approximate basis