一种关联规则增量更新算法

Incremental updating algorithm for mining association rules

针对一个新的事物数据库加到原有的事务数据库中,在最小支持度和最小置信度都不变的情况下相应关联规则的更新问题,提出了一种简单高效的增量式关联规则挖掘算法。该算法从集合的角度出发,将事物数据库分为频繁项集集合和非频繁项集集合,从而在两个事务数据库可能出现的关系中,准确找出能够生成新的频繁项集的集合。在此基础上,利用给出的算法能够较为容易地发现新的关联规则。最后,分析了此算法的优越性。

To update original association rules when a new transaction database is added to the original database and the minimum support threshold and minimum confidence threshold are kept unchanged, a simple and high-efficiency incremental updating algorithm for mining association rules is presented. Using the set method, a transaction database is divided into two sets, one is frequent itemset and the other is infrequent itemset. Accordingly, the set of frequent itemset can be discovered accurately in the possible relationship between two transaction databases. Based on this, a new association rule can be got easily by using the given algorithm. In the end, the superiority of this algorithm is analyzed.\;