Abstract: Mining of association rules is an important research topic among the various data mining problems, during which Apriori algorithm is a classic algorithm. And the most significant part of Apriori is the extraction of the frequent itemsets. This paper analyzes the performance of Apriori in mining association rules. To improve the connection step and pruning step of Apriori, a new algorithm named MApriori is introduced. In order to compare the two algorithms, an algorithm emluator is used. And the result indicates that the improved algorithm can accelerate the generating speed of the frequent multi-itemset, and improve the efficiency of mining accordingly.

Key words: Association rules, Frequent multi-itemset, Connection, Pruning

摘要： 关联规则挖掘是数据挖掘中的重要研究内容之一，Apriori算法是其中的经典算法，而频繁集的提取问题则是Apriori算法中的关键。该文对Apriori算法性能进行了分析，针对其中的连接步和剪枝步实施了改进，提出了MApriori算法。并通过算法仿真实验对这两种算法进行了比较，结果证明改进后的算法加快了高项频繁集的产生速度，从而提高了挖掘的效率。

关键词: 关联规则, 高项频繁集, 连接, 剪枝