电力调度数据挖掘后处理方法的研究  

Research on Post-processing of Data Mining for Electric Power Dispatching

作　　者：邹翔[1] 王清毅[1] 肖明军[1] 蔡庆生[1] 

机构地区：[1]中国科技大学计算机系,安徽合肥230027

出　　处：《小型微型计算机系统》2003年第6期1054-1057,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金 ( 60 0 75 0 15 )项目资助

摘　　要：关联规则是数据依赖关系的有效描述方法 ,是知识发现研究的重要内容 .然而 ,随着所挖掘数据库规模的增大 ,由传统数据挖掘算法所生成的大量关联规则常常令用户的使用与分析十分困难 .文中提出了一种新方法来解决这个问题并将其运用到电力调度数据挖掘系统中 .实验结果表明 ,该方法消除了大量冗余规则 ,并且使用户可以从整体上把握整个规则集 。Association rules is an effective method for describing the dependency relations in data, and it is one of the important aspects of knowledge discovery. However, as the size of mined databases increases,the sheer volume of rules produced by data mining often overwhelms their users. In this paper, we proposes a novel technique to overcome the problem and apply it in the electric power dispatching data mining system. Our experiments show that the method prunes a large number of redundant rules and makes user can obtain an overall picture of the whole domain, improving the accuracy and usability of association rule mining.

关 键 词：知识发现 数据挖掘 关联规则 代表规则 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]