| 基于相对可辨识矩阵的决策表属性约简算法 |
| |
| 引用本文： | 汪凌,吴洁,黄丹.基于相对可辨识矩阵的决策表属性约简算法[J].计算机工程与设计,2010,31(11). |
| |
| 作者姓名： | 汪凌 吴洁 黄丹 |
| |
| 作者单位： | 1. 西南交通大学经济管理学院,四川,成都,610031
2. 华东交通大学,理工学院,江西,南昌,330100 |
| |
| 基金项目： | 江西省教育厅科研基金 |
| |
| 摘 要： | 针对现有属性约简算法存在的问题,利用信息论和粗糙集理论,提出一种基于相对可辨识矩阵的决策表属性约简算法.该算法以核属性为基础,通过建立相对可辨识矩阵,利用条件信息熵作为启发式信息,减少属性约简过程中的搜索空间,逐个添加条件信息熵最大的属性,直到找出最小约简为止,并分析了该算法的时间复杂度.实例分析结果表明,该算法能有效地对决策表属性进行约简.
|
| 关 键 词： | 粗糙集 相对可辨识矩阵 信息熵 属性约简 启发式算法 |
Attribute reduction algorithm for decision table based on relative discernibility matrix |
| |
| WANG Ling,WU Jie,HUANG Dan.Attribute reduction algorithm for decision table based on relative discernibility matrix[J].Computer Engineering and Design,2010,31(11). |
| |
| Authors: | WANG Ling WU Jie HUANG Dan |
| |
| Affiliation: | WANG Ling1,WU Jie2,HUANG Dan2(1.School of Economics , Management,Southwest Jiaotong University,Chengdu 610031,China,2.Institute of Technology,East China Jiaotong University,Nanchang 330100,China) |
| |
| Abstract: | Based on information theory and rough set theory,a new attribute reduction algorithm based on relative discernibility matrix is proposed to solve the problems of the present attribute reduction algorithms.The core attribute is the foundation of this algorithm,through establishing the relative discernibility matrix to reduce the searching room,the condition entropy is used as heuristic information,and gradually add attributes which have the largest condition entropy,until getting the smallest reduction.And t... |
| |
| Keywords: | rough sets relative discernibility matrix information entropy attribute reduction heuristic algorithm |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
