基于FP-Tree的挖掘最大频繁项目集的新算法

A New Algorithm for Mining Maximum Frequent Itemsets Based on FP-Tree

挖掘最大频繁项目集是数据挖掘领域的一个重要的研究内容。Apriori算法作为一种挖掘频繁项目集的基本算法,其缺点是产生大量的候选项目集,算法的代价很大。本文在基于FP-Tree的基础上提出了挖掘最大频繁项目集的新算法FP-GDMA。该算法采用自顶向下和自底向上相结合的搜索策略有效减少了生产候选项目集的数目,有效提高了挖掘最大频繁项目集的效率。并通过实验比较FP-GDMA与DMFIA算法。

maximum frequent itemsets is an important research in data mining field. Apriori,as the basic algorithm for mining frequent itemsets,has drawbacks such as generating many candidate maximum frequent itemsets and tremendous cost .This paper proposes a new algorithm for mining maximum frequent itemsets named FP-GDMA based on FP-Tree. It uses the search strategy with combination of top down and bottom up,which effectively reduces the number of candidate frequent itemsets and effectively improves the efficiency for mining maximum frequent itemsets. The paper also compa, res FP--GDFM and DMFIA by exDeriment.