摘要： 关联挖掘算法中的Apriori 算法提供了一种根据查找频繁项集来发现数据集中的关联规则的方法，这种算法思路简单易于实现；但在由低次频繁项集生成高次频繁项集时需反复查找数据库，在效率上存在一定的欠缺，在寻找高次频繁项集时尤为明显。文章提出了一种新的关联规则挖掘算法：矩阵算法。同Apriori 算法相比较，该算法能直接查找高次频繁项集，可以有效地屏蔽Apriori 算法性能瓶颈。试验结果表明，当频繁项级较高时该算法比Apriori 具有更高的执行效率和性能，并具有良好的可行性。

关键词: 关联挖掘；Apriori 算法；频繁项集；矩阵算法

Abstract: Apriori algorithm can find out the associations of the data by finding the frequent itemsets by degrees. But it has the performance bottleneck when searching for the high level frequent itemsets. A new algorithm that can directly find the high level frequent itemsets is proposed in this paper. This algorithm can effectively resolve the bottleneck of Apriori. The result of the experiment shows that this algorithm can achieve better performance than Apriori and is more feasible especially when the degree of the frequent itemset is high.

Key words: Association mining; Apriori algorithm; Frequent itemset; Matrix algorithm