基于CAN-树的高效关联规则增量挖掘算法  

Efficient Incremental Association Rules Mining Algorithm Based on CAN-tree

作　　者：邹力鹍[1] 张其善[1] 

机构地区：[1]北京航天航空大学电子信息工程学院,北京100083

出　　处：《计算机工程》2008年第3期29-31,共3页Computer Engineering

摘　　要：关联规则是数据挖掘领域的一个重要研究方向。针对关联规则的增量挖掘问题,该文提出一种快速算法FIAFAR。算法使用CAN-树存储原始交易数据库,弥补了FP-树的不足,适应于增量挖掘以及最小支持度变化的情况。采用子父节点指针的设计,可以快速生成条件模式树,提高算法的效率。实验验证了算法的有效性。Finding association rules is a major aspect of data mining research. An efficient algorithm FIAFAR is proposed to deal with incremental mining of association rules. Algorithm is based on the CAN-tree structure, overcomes the shortcomings of FP-tree which it is not suitable for incremental mining and minimum support varying. The children to parent pointer is used in algorithm instead of parent to children pointer. It helps algorithm FIAFAR build conditional pattern tree effectively and improves the efficiency of algorithm in consequence. Experimental results show that the new algorithm is effective for large databases.

关 键 词：数据挖掘 关联规则 条件模式树 指针 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]