引用本文:许孝元, 韩国强, 闵华清.多步原子规则的大规模关联分类[J].控制理论与应用,2007,24(3):471~474.[点击复制]
XU Xiao-yuan, HAN Guo-qiang, MIN Hua-qing.Multistep classification based on atomic and associative rules in the large-scale datasets[J].Control Theory and Technology,2007,24(3):471~474.[点击复制]
多步原子规则的大规模关联分类
Multistep classification based on atomic and associative rules in the large-scale datasets
摘要点击 1460  全文点击 1215  投稿时间:2004-10-25  修订日期:2007-01-15
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/j.issn.1000-8152.2007.3.027
  2007,24(3):471-474
中文关键词  数据挖掘  机器学习  关联规则  分类
英文关键词  data mining  machine learning  association rules  classification
基金项目  广东省科技攻关资助项目(2003C101007); 广州市科技计划基金资助项目(2004Z3-E0091)
作者单位
许孝元, 韩国强, 闵华清 华南理工大学计算机科学与工程学院, 广东广州510640
广东工业大学计算机学院, 广东广州510090) 
中文摘要
      关联分类已成为数据挖掘研究的热点问题之一. 为解决大规则关联分类问题, 本文提出了基于原子规则的多步分类方案, 并对作者提出的多步原子关联规则分类新技术进行了深入的理论研究. 与同类关联分类方法(如CBA)比较, 本文提出的方法具有学习速度快、分类准确度高的优点.
英文摘要
      The principles of multi-step classification based on atomic rules are described in this paper. A related scheme based on atomic rules is proposed to tackle the problem of associative classification in the context of large-scale datasets. Compared to the well-known associative classification algorithm of CBA, the proposed approach has the advantage of fast speed and high accuracy.