关键词: 频繁项集, 矩阵, 二进制, 前缀树

Abstract: Traditional algorithm of frequent itemsets mining’s execution efficiency is low. MPFI algorithm based on matrix and Prefix-tree is raised for frequent itemsets mining. It can quickly mining frequent itemsets of transaction database. MPFI algorithm only scan transaction database once, builds vertical binary matrix. Binary vector for frequent itemsets information and Prefix-tree data structure for compression storage of frequent itemsets is applied in the algorithm, without candidate itemsets. According to theoretical analysis and experimental results, MPFI algorithm can effectively improve the efficiency of frequent itemsets mining.

Key words: frequent itemsets, matrix, binary, Prefix-tree