An Efficient Algorithm for Discovering all Frequent Patterns

Discovering frequent patterns is one of the essential topic data mining. An efficient algorithm based on compressed bitmap and two-way-hybrid strategy is proposed. Firstly, a hierarchical search space organization is presented, based on which the original search space can be recursively decomposed into some smaller independent pieces. Secondly, a compressed bitmap index technique is employed to speed up the support counting. The original bitmap can be impacted efficiently by this way. Thirdly, a novel HFMI algorithm, which explores a flexible two-way-hybrid search method, is given. It can discovery both longer and shorter frequent patterns as many as possible. And experimental and analytical results are presented in the end of this paper.