中文摘要：

Generator项集是频繁项集的一种简洁表示形式,但当为数据项附以不同权值时,传统Generator定义和计算方法将不再有效;同时其具有的反单调性也不再成立;而此特性在现实应用中具有重要作用。提出以设计权重约束下的支持度计算方法为基础,重新定义加权Generator,调整加权Generator与频繁项集的计算关系,从而维持其反单调性。在加权支持度树结构的基础上,给出挖掘加权Generator的方法。实验表明,提出的方法能够使得权重较大的项被优先挖掘出来,并保证了Genera-tor的反单调性,提出的挖掘方法能够正确有效挖掘频繁加权Generator。

英文摘要：

Generator is a concise representation for frequent itemsets.But in real circumstances,when items are attached with weights,general definition of generator may come into false.And its properties don’t hold,for example anti-monotone property which is useful in real application.Generator with weight is defined through tuning its support calculation according to weights.And generator with weight keeps the anti-monotone property.Based on weighted support tree structure,a new method is proposed for mining all frequent generators with weights.Experimental results show that the items with larger weights are more likely derived than those with smaller ones.And the proposed method can mine all frequent generator with weights effectively.