| 基于加权综合的样本分类算法 |
| |
| 引用本文： | 刘文军,杨培亮.基于加权综合的样本分类算法[J].大理学院学报,2006,5(10):5-7. |
| |
| 作者姓名： | 刘文军 杨培亮 |
| |
| 作者单位： | 1. 长沙理工大学数学与计算科学学院,湖南,长沙,410076
2. 大理学院数学与计算机学院,云南大理,671003 |
| |
| 基金项目： | 长沙理工大学校科研和教改项目 |
| |
| 摘 要： | 首先分析了粗糙集理论中现有属性依赖性定义的不合理性，然后给出一种新的属性依赖性度量．根据这种依赖性度量，给出属性重要性的定义，再以这种属性重要性为权重，给出一种基于加权综合的样本分类算法。由于属性的重要性是由条件属性相对于决策属性的依赖性决定的，它的数值相对比较客观，这样解决了常用的定权方法的弱点，它们一般是凭经验或由专家给出，具有相当的主观性。
|
| 关 键 词： | 粗糙集 依赖度 重要性 加权 |
| 文章编号： | 1672-2345(2006)10-0005-03 |
| 收稿时间： | 04 7 2006 12:00AM |
| 修稿时间： | 2006年4月7日 |
The Classification Algorithm Based on Weighed Comprehensive |
| |
| LIU Wen-jun,YANG Pei-liang.The Classification Algorithm Based on Weighed Comprehensive[J].Journal of Dali University,2006,5(10):5-7. |
| |
| Authors: | LIU Wen-jun YANG Pei-liang |
| |
| Institution: | 1, Department of Mathematics and Computing Science,Changsha University of Science and Technology,Changsha Hunan 410076, China; 2, Department of Mathematics and Computer, Dali University, Dali, Yunnan 671003,China |
| |
| Abstract: | First, we analyze the unreasonableness of the definition of dependency degree of attributes in rough sets, then give a new measure of degree of dependency. According to this degree of dependency, we give a definition of degree of attribute significance, Then a weighed comprehensive classification algorithm is given. Because the significance of attribute is computed on the basis of dependency between condition and decision attribute, its value is objective. This method compensates the drawback that it is difficult to give significance of a attribute, |
| |
| Keywords: | rough sets dependency degree significance weigb |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
