关键词: 最大频繁项集 多层扩展 深度优先搜索 前瞻剪枝

Abstract:

We present KMAX, a new depth-first search algorithm for mining maximal frequent itemsets. KMAX uses a novel technique called multilevel extension to extend the items in the search tree with an efficient look-ahead pruning method to prune the search space. Experimental comparison with the previous work indicates that it obviously accelerates the generation of maximal frequent itemsets , therefore the CPU time is reduced remarkably.

Key words: maximal frequent itemset, multilevel extension, depth-first search, look-ahead pruning