Mining frequent closed itemsets using antecedent-consequent constraint and length-decreasing support constraint

At present, many frequent itemsets mining algorithms adopt a constant support threshold value strategy, which is not convenient to potential valuable long itemsets discovery. Length-decreasing support constraints can address this problem probably. Existing algorithms make improvements on classical algorithm, which result to low efficiency. Tailored to the medical data, this paper proposes a frequent closed itemsets mining algorithm called ACLCMiner, which uses the antecedent-consequent constraint and the length-decreasing support constraint. With the antecedent-consequent constraint, the algorithm greatly reduces the number of generated frequent itemsets. The experimental results show that ACLCMiner is efficient and can find more long itemsets with potential values.