摘要
关联规则通常以规则列表形式表达,而许多关联规则挖掘算法往往产生大量规则,这给用户理解规则和从中找出感兴趣的规则带来了极大困难。为了标识重要的规则,而又保持挖掘结果的完整性,提出了根据规则的通用性,按照由概括—具体的方式分层表达关联规则。先用挖掘结果的最概括规则集表达出最通用、最基本的领域知识,再根据用户要求分层查看概括规则下面更具体的规则。这种表达方式可以在不同层次上查看关联规则,使挖掘结果更容易管理和被人理解。
Association rules are always presented as a list of rules with no further organization. Because many existing associations mining algorithms often produce a large number of rules, it is very difficult for the user to analyze them manually. In order to highlight the important rules and not suffer from being incomplete, a technique was proposed to intuitively present the discovered association rules in a hierarchical fashion. The discovered association rules were shown at different levels according rule's generality. The technique first found a subset of the association rules called the most-general rules set to give the user a general relationship or a big picture of the original rules set. Then, the user could selectively view more specific rules bellow the general rule that were interesting to him/her. Presented in such a way, rules could be easily comprehended by users.
出处
《计算机应用》
CSCD
北大核心
2006年第1期207-209,共3页
journal of Computer Applications
关键词
数据挖掘
关联规则
层次
概括规则
具体规则
data mining
association rules
hierarchy
more-general rules
more-specific rules