Algorithms for Mining Frequent Itemsets with Multi-Predication Constraints Based on Frequent Pattern Growth

Aiming at mining multidimensional frequent itemsets from affair database,the conception of mining with multidimensional constrained is brought forward.Two algorithms are designed according to FP-growth and predication constraint,the MCMFI1 must construct FP-tree for every constraint as another algorithm MCMFI2 which based on node vector constrained have more excellent performance in searching and updating the existing FP-Tree and itemsets,but spending added memory.The analyses and experiments prove algorithms are effective.