摘要
该文提出了基于聚类的比例规则(CRR),该规则保留了比例规则(RR)的优良特性,同时适用于多峰分布的数据集。实验证明,基于聚类的比例规则的预测效果优于比例规则的预测效果。
Ratio Rules based on clusters are proposed in this paper.They not only reserve the fine features of Ratio Rules,but also suit datasets which have many peaks.Experiments on two real datasets demonstrate that the proposed method achieves less root-mean-square than Ratio Rules.
出处
《计算机工程与应用》
CSCD
北大核心
2001年第14期115-117,173,共4页
Computer Engineering and Applications
基金
云南省自然科学基金资助
关键词
比例规则
数据挖掘
预测
基于聚类的比例规则
Ratio Rules(RR), Data Mining, Forecast, Ratio Rules based on clusters(CRR)