Abstract: Enhancing the efficiency of frequent itemset mining algorithm is an important area of researching association rule mining. This paper presents an array-based association rules mining algorithm, which only scans database once and progressively reduces the number of transactions in database. Moreover, it adopts one-dimension array to count the number of candidate 2-itemsets for expediting the mining process. Experiments show that the algorithm is faster than the Apriori by a factor from two to three.

摘要： 提高频繁项集挖掘算法的效率是关联规则挖掘研究的一个重点领域。文章提出了基于数组的关联规则挖掘算法，只需要扫描数据库1次，通过不断减少数据库中的事务个数，并且利用一维数组对候选2-项集进行计数来提高挖掘效率。实验表明，该文所提出的算法效率比经典Apriori算法快2~3倍。

关键词: 据挖掘, 关联规则, 频繁项集

CLC Number: 

• TP311.13