自上而下的关联规则采掘方法  

Mining Association Rules Using Top-down Method

作　　者：何嵩[1] 田振军[1] 

机构地区：[1]复旦大学计算机科学系,上海200433

出　　处：《计算机科学》1999年第1期62-64,共3页Computer Science

摘　　要：1.前言在数据集中采掘关联规则是数据采掘的一个重要内容,[2]中提出了一种自底向上的数据采掘方法,即把数据集的各个属性从概念层次的最底层向高层进行“泛化”,但这种方法有二个缺点:首先,它的每一层都有一个阈值,但这些阈值之间是密切相关的;并且,对于不同数据集,这些阈值之间的关系也不同。在数据采掘的过程中。Mining association rules from a data-set is an important branch in data mining area. In a bottom-up method, we have to adjust the thresholds at different levels many times for good effect, and the process of adjustment is hard to control. On the other hand, the bottom-up method can only assign field-names to the values at the same level. In this paper, we first raised some definitions about the specialization and also some theorems. Then we gave our top-down method of the association rule mining on this base. This method can effectively resolve these two problems with the same good efficiency.

关 键 词：关系数据库 关联规则 数据采掘 数据集 数据库 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]