| 基于粗糙集的RDT决策树生成算法的研究及应用 |
| |
| 引用本文： | 江效尧,胡林生.基于粗糙集的RDT决策树生成算法的研究及应用[J].安徽工程科技学院学报,2004,19(3):50-53. |
| |
| 作者姓名： | 江效尧 胡林生 |
| |
| 作者单位： | 南京审计学院,计算机系,江苏,南京,210029 |
| |
| 基金项目： | 南京审计学院校科研和教改项目 |
| |
| 摘 要： | 介绍了一种基于粗糙集理论的决策树生成算法--RDT(Rought Set Decision Tree).该方法运用了粗糙集理论中条件属性相对于决策属性的核,引入启发式条件计算并选择条件属性作为决策树的根结点或子结点.通过一个例子,与运用信息熵概念建立决策树的算法进行比较,结果表明采用RDT方法得到的决策树优于采用信息熵方法得到的决策树.还讨论了RDT与ID3算法对决策树精度和规模的影响,分析数据分类和知识发现的过程及特点.
|
| 关 键 词： | 粗糙集 决策树 分类 知识发现 |
| 文章编号： | 1672-2477(2004)03-0050-04 |
| 修稿时间： | 2004年4月10日 |
Reasearch and application of the RDT algorithm based on decision tree in data mining |
| |
| JIANG Xiao-yao,HU Lin-sheng.Reasearch and application of the RDT algorithm based on decision tree in data mining[J].Journal of Anhui University of Technology and Science,2004,19(3):50-53. |
| |
| Authors: | JIANG Xiao-yao HU Lin-sheng |
| |
| Abstract: | This paper introduces how to produce a decision tree algorithm-RDT(Rough Set Decision Tree) which is based on Rough Set. The method adopts the core of condition attributes with respect to decision attributes, and calculates condition of heuristic to find root or root of subtree. Comparing decision tree produced by RDT with ID_3 algorithm, the results indicate that decision tree produced by RDT is better than decision tree produced by information entropy. Besides the precision of classification and the scale of the tree, the analysis of the processes and characters of data classification and knowledge discovery are also reflected in this paper |
| |
| Keywords: | rough set decision tree classification knowledge discovery |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
