Abstract: This paper proposes a depth-first algorithm for mining generator representation. Through candidates pruning based on suffix information, it avoids much redundant work and needs not to frequently scan original database. Experimental results show it is an effective method for generator representation mining.

Key words: Association rule, Data mining, Generator

摘要： Generator表示方法使得基于频繁集的挖掘任务更加简洁高效。该文提出深度优先挖掘Generator表示的方法，通过候选剪枝和在树结构上通过保留挖掘后缀信息，完成Generator候选检测，而不需重复扫描数据库。同时采用的剪枝策略使挖掘冗余操作尽可能减少，取得了较好的效果。实验表明提出的算法是一种有效挖掘Generator表示的方法。

关键词: 关联规则, 数据挖掘, Generator