Frequent Itemset Mining Algorithm Based on Graph and Two-directional Search
Fang Liu · Jisuanji gongcheng · 2012
For discovering association rules based on graph can generate a large number of candidate itemsets,an improved algorithm is proposed.The improved algorithm combined the top-down and bottom-up search in all search process,and sorted the frequent 1-itemset on support degree and count the support of maximal superset of frequent k-itemsets.It utilizes direct graph and Apriori property to prune the redundant candidate itemsets.Experimental result shows that the improved algorithm reduce the number of candidate itemsets when the minimum support is small and the performance is improved.