关键词: 频繁项目集, 支持度, 信任度, 概念格, 等价关系

Abstract: Among some association algorithms, Apriori algorithm should be made better in processing item and transaction sets, for example, its efficiency in creating frequent 2-item L2 is very low. Apriori algorithm was researched and its merits and defects were analysed. A new association rule algorithm, Apriori algorithm based on concept lattices (ACL), was put forward. It adopted concept lattices and equivalence relationship, made use of the theory of rough sets, and then calculated to gain frequent 2-item L_2. Experiment results show that ACL is an efficient algorithm, and outperforms previous methods.

Key words: frequent itemset, support, confidence, concept lattice, equivalence relationship