基于频繁项集特性的Apriori算法的改进

Improvement of Apriori algorithm based on characteristic of frequent itemset

Apriori算法是关联规则中一种重要算法。Apriori算法在求出频繁项集的过程中,需要扫描事务项集里的数据。由于事务项集里只是部分数据有用,所以改进算法,缩小所需扫描的事务项集大小,并提出了一种简单的数据结构——树型结构来存储事务项集数据,使得算法在数据集量巨大时,性能得到有效提高,并用实例验证了这些改进能够正确、有效、快速地实现该算法。

Apriori algorithm is one of important algorithms in association rule algorithms. The process of get frequent itemset in Apriori algorithm needs scan every data in transaction itemset, In fact, only part of data in transaction itemset are availability, The improvement of apriori algorithm here is to decrease transaction itemset data, and created a tree structure which used to store data sets. To huge scale data sets, the performance of algorithm improved effectively. An example is cited to prove these improvements.