| 基于类间可分性度量的支持向量机决策树 |
| |
| 引用本文： | 韩虎,任恩恩.基于类间可分性度量的支持向量机决策树[J].计算机工程与设计,2007,28(18):4454-4455,4458. |
| |
| 作者姓名： | 韩虎 任恩恩 |
| |
| 作者单位： | 1. 兰州交通大学,数理与软件工程学院,甘肃,兰州,730070;兰州交通大学,光电技术与智能控制教育部重点实验室,甘肃,兰州,730070
2. 兰州交通大学,光电技术与智能控制教育部重点实验室,甘肃,兰州,730070 |
| |
| 摘 要： | 采用支持向量机解决多类分类问题一般通过多个两类分类器的组合来求解,如何组合这些两类分类器就是该方法的关键.提出一种改进的支持向量机决策树多类分类模型,该模型通过引入类间可分性度量来确定决策树结构,以类间可分性度量的高低来决定不同类别在决策树中的位置,将容易分离的类尽可能早地划分出来.最后通过一组实验证明了该模型的有效性.
|
| 关 键 词： | 支持向量机 多类分类 组合 决策树 类间可分性 可分性度量 支持向量机 决策树 separability based decision tree 有效性 验证 划分 分离 位置 类别 树结构 分类模型 改进 方法 求解 组合 分类器 多类分类问题 |
| 文章编号： | 1000-7024（2007）18-4454-02 |
| 修稿时间： | 2006-10-24 |
Classifiable model with SVM decision tree based on inter-class separability |
| |
| HAN Hu,REN En-en.Classifiable model with SVM decision tree based on inter-class separability[J].Computer Engineering and Design,2007,28(18):4454-4455,4458. |
| |
| Authors: | HAN Hu REN En-en |
| |
| Affiliation: | 1. College of Mathematics, Physics and Software Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China; 2.Key Laboratory Opto-Electronic Technology and Intelligent Control,Lanzhou Jiaotong University, Lanzhou 730070 ,China |
| |
| Abstract: | Solving a multi-class classification problem with SVM is usually converted to several binary classification problems, how to combine these binary classification is important for classifiable efficiency. An improved SVM decision tree is presented, the structure of the SVM decision tree is defined based on inter-class separability measure, and these classes with higher separability are separated as early as possible. The experimental results show the performance of the method. |
| |
| Keywords: | support vector machine multi-class classification combine decision tree inter-class separability |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
