首页 | 本学科首页   官方微博 | 高级检索  
     

基于二进制可辨矩阵的属性约简算法的改进
引用本文:周海岩,杨汀.基于二进制可辨矩阵的属性约简算法的改进[J].计算机工程与设计,2003,24(12):35-37,42.
作者姓名:周海岩  杨汀
作者单位:太原师范学院,计算机系,山西,太原,030012
基金项目:山西省高校科技研究开发基金资助项目(Z002056)
摘    要:属性约简是粗糙集理论的核心内容之一,信息系统中知识(属性)并不是同等重要的,甚至其中某些知识是冗余的。属性约简是在保持信息系统中知识量(即分辨能力)不变的条件下,删除其中不相关或不重要的知识,现已证明寻找信息系统的最小属性约简是NP-hard问题。解决这类问题的一般方法是采用启发式算法求出最优或次最优约简。对支天云等所给出的二进制可辨矩阵的化简算法进行了改进,并根据属性的分辨能力的大小,提出了一种基于二进制可辨矩阵的思路清晰、实现简便的属性约简算法。通过算法分析表明,该算法是更加高效的。

关 键 词:粗糙集理论  属性约简算法  知识发现  二进制可辨矩阵  信息系统  人工智能
文章编号:1000-7024(2003)12-0035-03

A more efficient algorithm for attribute reduction based on the binary discernible matrix
ZHOU Hai - yan,YANG Ting.A more efficient algorithm for attribute reduction based on the binary discernible matrix[J].Computer Engineering and Design,2003,24(12):35-37,42.
Authors:ZHOU Hai - yan  YANG Ting
Abstract:Attribute reduction is one of the key problems for the rough set theory. It is not equally important in attributes of the information system, even some are redundant. Some information which are unimportant or non-relative can be deleted in the case that the resolutions of the information are not be changed. Nowadays the NP-hard problem has been proposed for searching the minimum attribute reduction of the information system. The common method to solve this problem is to adopt the heuristic algorithm. According to Zhi Tian-yun's binary discernible matrix reducible algorithm and the algorithm for attribute reduction based on the binary discernible matrix, the author designed a more efficient algorithm.
Keywords:rough sets  binary discernible matrix  reduction of attribute
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号