应用科学学报 ›› 2005, Vol. 23 ›› Issue (6): 615-619.

• 论文 • 上一篇    下一篇

多分类-关联规则分类的一种改进算法

赵传申, 孙志挥   

  1. 东南大学计算机科学与工程系, 江苏南京 210096
  • 收稿日期:2004-08-11 修回日期:2004-10-27 出版日期:2005-11-30 发布日期:2005-11-30
  • 作者简介:赵传申(1973-),男,山东高唐人,博士生,E-mail:zhao_chuanshen@dalutech.com;孙志挥(1941-),男,江苏南通人,教授,博导.
  • 基金资助:
    国家自然科学基金资助项目(70371015)

An Improved Classification Algorithm Based on Multiple Class-Association Rules

ZHAO Chuan-shen, SUN Zhi-hui   

  1. Department of Computer Science and Engineering, Southeast University, Nanjing 210096, China
  • Received:2004-08-11 Revised:2004-10-27 Online:2005-11-30 Published:2005-11-30

摘要: 提出了一种基于多分类-关联规则的快速分类算法——FCMAR,该算法在建立频繁模式树(FP-tree)时裁减掉不能生成频繁规则的项目,因而可减少FP-tree的节点数目,有效地降低时间和空间复杂度,实验结果表明该算法是有效可行的.

关键词: 数据挖掘, 分类, 关联分类, 频繁模式树

Abstract: Classification is an important aspect in data mining research.In this paper, a new associative classification algorithm, fast classification based on multiple association rules (FCMAR), is proposed.This algorithm removes items that do not generate frequent rules so that storage space is saved and efficiency improved.Experimental results show that the FCMAR algorithm is efficient and effective.

Key words: classification, frequent pattern tree (FP-tree), associative classification, data mining

中图分类号: