计算机应用 ›› 2012, Vol. 32 ›› Issue (02): 304-308.DOI: 10.3724/SP.J.1087.2012.00304

• 数据库技术 • 上一篇    下一篇

基于显露模式的对比挖掘研究及应用进展

段磊1,唐常杰1,Guozhu Dong2,杨宁1,苟驰1   

  1. 1. 四川大学 计算机学院,成都 610065
    2. Wright State University, Department of Computer Science and Engineering, Dayton 45435, USA
  • 收稿日期:2011-09-22 修回日期:2011-10-31 发布日期:2012-02-23 出版日期:2012-02-01
  • 通讯作者: 唐常杰
  • 作者简介:段磊(1981-),男,四川成都人,讲师,博士,主要研究方向:数据挖掘;
    唐常杰(1946-),男,重庆人,教授,博士生导师,博士,主要研究方向:数据库、知识工程;
    Guozhu Dong(1957-),男,美籍华人,教授,博士生导师,主要研究方向:数据挖掘;
    杨宁(1974-),男,四川成都人,讲师,博士,主要研究方向:数据挖掘;
    苟驰(1990-),男,四川泸州人,主要研究方向:数据挖掘。
  • 基金资助:
    国家自然科学基金资助项目(61103042,61173099);高等学校博士学科点专项科研基金资助项目(20100181120029);四川大学青年教师科研启动基金资助项目(2009SCU11030)

Survey on emerging pattern based contrast mining and applications

DUAN Lei1,TANG Chang-jie1,Guozhu DONG2,YANG Ning1,GOU Chi1   

  1. 1. College of Computer Science, Sichuan University, Chengdu Sichuan 610065, China
    2. Wright State University, Department of Computer Science and Engineering, Dayton 45435, USA
  • Received:2011-09-22 Revised:2011-10-31 Online:2012-02-23 Published:2012-02-01
  • Contact: TANG Chang-jie

摘要: 对比挖掘是近年来数据挖掘领域的新热点之一。对比挖掘关注并描述不同类别和条件下,或随时间变化的知识,旨在设计能够发现刻画数据集中不同类别或条件的样本间差异的模式或模型的方法。由于对比挖掘技术能化繁为简、准确分类,在实践中得到广泛应用。显露模式的挖掘和应用是对比挖掘的重要分支。综述了显露模式的背景、基本概念和原理,分析了显露模式的挖掘方法,讨论了显露模式的扩展定义和挖掘,介绍了基于显露模式的分类器构造方法,展示了显露模式的若干实际应用,展望了基于显露模式的对比挖掘的未来研究。

关键词: 数据挖掘, 显露模式, 模式发现, 频繁项集, 分类

Abstract: Contrast mining is one of fairly new hot data mining topics. Contrast mining focuses on knowledge that describes differences between classes and conditions, or describes changes over time. Contrast mining aims at developing techniques to discover patterns or models that contrast, and characterize multiple datasets associated with different classes or conditions. Contrast mining has wide applications in reality, due to its ability of simplifying problems and classifying accurately. Research on the mining and application of emerging patterns represents a major direction of contrast mining. This paper provided a survey of such issue. More specifically, after introducing the background, basic concepts and principles of emerging patterns, the paper analyzed the mining methods of emerging patterns, discussed extended definitions of emerging patterns and their mining, stated methods for constructing emerging pattern based classifiers, and illustrated applications of emerging pattern in several real-world fields. Finally, this paper gave out some topics for future research on emerging pattern based contrast mining.

Key words: data mining, emerging pattern, pattern discovery, frequent itemset, classification

中图分类号: