首页 | 本学科首页   官方微博 | 高级检索  
     

决策表化简及其属性约简方法
引用本文:桂现才. 决策表化简及其属性约简方法[J]. 计算机工程与设计, 2007, 28(8): 1765-1767
作者姓名:桂现才
作者单位:湛江师范学院,数学与计算科学学院,广东,湛江,524048
基金项目:湛江师范学院校科研和教改项目
摘    要:给出一种求不一致决策表的核和所有属性约简的新方法,首先利用U/C对决策表进行化简,然后在简化的决策表上定义新的差别矩阵,最后利用差别函数求出所有属性约简.该方法可有效降低生成差别矩阵时所需的存储空间和时间,理论分析和例子表明,该属性约简方法在效率上比现有的基于差别矩阵和差别函数的方法有显著提高.

关 键 词:属性约简    正区域  差别矩阵  差别函数  决策表
文章编号:1000-7024(2007)08-1765-03
修稿时间:2006-04-10

Simplified decision table and its attributes reduction
GUI Xian-cai. Simplified decision table and its attributes reduction[J]. Computer Engineering and Design, 2007, 28(8): 1765-1767
Authors:GUI Xian-cai
Affiliation:School of Mathematics and Computational Science, Zhanjiang Normal College, Zhanjiang 524048, China
Abstract:A new method for computing core and attributes reduction of the inconsistent decision table is presented. First, U/C is used to simplify decision table. Furthermore, a new discernibility matrix is defined in simplified decision table. Finally, discernibility function is used to computing all attributes reductions. The method is reduced more stored spaces and times in produced discernibility matrix. Theoretical analysis and example show that the attributes reduction method is more efficient than those existing methods based on discernibility matrix and discernibility function.
Keywords:attribute reduction   core   positive region   discernibility matrix   discernibility function   decision table
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