摘要
详细讨论了用于关联规则可视化的几种常用技术,分析了各自的优缺点。提出了关联规则可视化的一种新方法ARVir,该方法巧妙地利用平行坐标技术的思想,对原有的可视化技术进行改进,能够解决当前关联规则可视化技术中普遍存在的界面紊乱、产生歧义等多种问题。利用Java3D技术实现了基于ARVir的关联规则可视化系统原型,实验表明该系统不仅能够有效地显示大量关联规则,而且用户可以给定约束条件对挖掘结果进行过滤。
This paper discusses several useful techniques which are usually used to visualize association rules with the advantages and disadvantage of each technique presented, and then introduces a novel method ARVir which ingeniously improving parallel coordinates technology to visualize association rules. The method ARVir is an improvement over those existed visualization techniques and can tackle many of common problems such as occlusion and confusion in the visualizing process. A system for visualizing association rules based ARVir is implemented with Java3D. The system not only efficiently visualizes abundant association rules, but also filters the mined results by constraint condition.
出处
《计算机工程》
EI
CAS
CSCD
北大核心
2005年第24期87-89,共3页
Computer Engineering
基金
江苏省自然科学基金资助项目(BK2005135)
江苏省重点实验室开放基金资助项目(KJS03064)
关键词
关联规则
可视化
数据挖掘
Association rule
Visualization
Data mining