一种约束最大频繁模式快速挖掘算法

Fast Algorithm for mining constrained maximal frequent patterns

为了避免用户通过"二次挖掘"才能得到有用的结果集,本文提出了一种新的约束最大频繁模式挖掘算法CSMFP-Max。CSMFP-Max算法基于CFP树和对称矩阵,在挖掘过程中采用了多种剪枝策略并结合了自顶向下和自底向上的双向搜索策略,大大缩小了候选集规模,避免了不必要的条件CFP树的产生。理论分析和实验结果表明CSMFP-Max算法是一种高效的约束最大频繁模式挖掘算法,具有良好的时空效率。

This thesis proposes a new algorithm CSMFP-Max for mining constrained maximal frequent patterns to avoid users using second time to get useful results. And during the mining process, this algorithm based on CFP-tree and symmetrical matrix introduces many pruning strategies and a search strategy that integrating the top to down and down to top method, which reduce the size of candidates and avoid generating numerous unnecessary term CFP-trees. Theoretical analysis and experimental results show that CSMFP- Max is a high performance algorithm for mining constrained maximal frequent patterns. It has better efficiency in time and space.