基于规则约束的并行FP-Growth算法研究

Parallel FP-Growth Algorithm Based on Constraint Rule

频繁模式增长(FP-Growth)算法是一种以发现频繁项集为基础的关联规则挖掘算法,从实际应用中发现,该算法需要挖掘出全部频繁项集,导致挖掘效率不高,并且无法适应大数据挖掘。因此,在现有研究的基础上,为适应大数据挖掘,进一步提高该算法的效率,论文针对该算法存在的不足,提出一种基于规则约束的并行FP-Growth算法,即在并行计算模式(SIMD-SM)下对挖掘对象进行规则约束。

Frequent-Pattern Growth Algorithm（FP-Growth） is an association rules mining algorithm based on finding frequent itemsets .According to findings from actual applications ,this algorothm need to find out all frequent itemsets .On this account the mining efficiency becomes low and it can not accommodate big date mining .So ,on the basis of existing re-search ,a parallel FP-Growth algorithm is presented to accommodate big date mining and improve the efficiency of original al-gorithm .Parallel computing model（SIMD-SM ） and constraint rules are adopted in the new algorithm .The new algorithm can find out all frequent itemsets and can deal with mass data very well .