FpGrowth algorithm-based pruning strategy

An FpGrowth algorithm is an association rule mining algorithm which does not generate candidate sets, and has a wide practical application value. Through deeply analyzing and researching a structure and a mining process of an FP-tree of a classic FpGrowth algorithm, and analyzing a single-path mining method and a multi-path mining method of the FP-tree, the invention discloses a pruning strategy which is capable of decreasing the iteration frequencies of a part of branches. Experiment results prove that the pruning strategy is capable of effectively improving the algorithm and improving the data processing ability and efficiency of the FpGrowth algorithm. FpGrowth算法是不产生候选集的关联规则挖掘算法，具有广泛的实际应用价值。为此对经典FpGrowth算法的FP-tree的结构和挖掘过程进行了深入分析和研究，分析了FP-tree单路径和多路径的挖掘方法，提出了个剪枝策略，在频繁模式挖掘时可以减少部分分支的迭代次数。实验结果验证提出的剪枝策略有效地改进了算法，并提高了FpGrowth算法对数据的处理能力和效率。