Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (9): 150-151.

• 数据库、信号与信息处理 • Previous Articles     Next Articles

Study of reduction algorithms based on discernibility matrix of length constraint

WANG Hao1,2,HU Xue-gang1   

  1. 1.School of Computer and Information,Hefei University of Technology,Hefei 230009,China
    2.Network Information Center,Anhui Institute of Architecture & Industry,Hefei 230022,China
  • Received:2007-07-10 Revised:2007-09-17 Online:2008-03-21 Published:2008-03-21
  • Contact: WANG Hao

基于长度约束区分矩阵的约简算法研究

王 浩1,2,胡学钢1   

  1. 1.合肥工业大学 计算机与信息学院,合肥 230009
    2.安徽建筑工业学院 网络信息中心,合肥 230022
  • 通讯作者: 王 浩

Abstract: Rough set theory is one of the main subjects in the field of machine learning and data mining.The application of this theory is mainly realized through attribute reduction algorithms.This paper presents a reduction algorithm based on discernibility matrix of length constraint(RABDMLC).By counting the average length of discernibility matrix item in the sampling dataset and deleting the lenght in the process of constructing discernibility matrix,the efficiency of reduction is thus improved.It is demonstrated that RABDMLC is effective and feasible through the contrastive experiment between RABDMLC and attribute reduction algorithm based on attributes frequency(ARABAF).

Key words: rough set, sampling, length constraint, discernibility matrix, attributes reduction

摘要: 粗糙集理论是机器学习和数据挖掘领域的重要课题之一,其中属性约简算法是该理论实现应用的主要算法。提出了一种基于长度约束区分矩阵的约简算法(RABDMLC算法),通过抽样数据集计算平均区分矩阵项长,构造区分矩阵时不构造长于平均区分矩阵项长的项,在一定程度上提高了约简的效率。与基于属性频度函数的约简算法进行对比试验分析后,验证了该算法是有效和可行的。

关键词: 粗糙集, 抽样, 长度约束, 区分矩阵, 属性约简