| 基于启发式信息的支持向量机规则抽取 |
| |
| 引用本文： | 张德贤,张苗,谭一鸣.基于启发式信息的支持向量机规则抽取[J].计算机应用,2008,28(3):729-731. |
| |
| 作者姓名： | 张德贤 张苗 谭一鸣 |
| |
| 作者单位： | 河南工业大学,信息科学与工程学院,郑州,450001 |
| |
| 摘 要： | 提出一种基于启发式信息的支持向量机规则抽取方法，解决了规则抽取中处理连续属性困难的问题。将支持向量回归（SVR）分类超曲面位置和形状特征作为启发式信息指导属性选择和属性区间的合理分割，然后给出了分类规则抽取的具体算法。通过UCI中多个数据集进行测试，证明与其他规则抽取方法相比，该方法显著提高了分类规则抽取的效率，尤其对复杂的分类问题。
|
| 关 键 词： | 启发式信息 支持向量机 支持向量回归 规则抽取 |
| 文章编号： | 1001-9081（2008）03-0729-03 |
| 收稿时间： | 2007-09-14 |
| 修稿时间： | 2007-11-27 |
Extracting symbolic rules from support vector machines based on the heuristic information |
| |
| ZHANG De-xian,ZHANG Miao,TAN Yi-ming.Extracting symbolic rules from support vector machines based on the heuristic information[J].journal of Computer Applications,2008,28(3):729-731. |
| |
| Authors: | ZHANG De-xian ZHANG Miao TAN Yi-ming |
| |
| Affiliation: | ZHANG De-xian,ZHANG Miao,TAN Yi-ming(School of Information Science , Engineering,Henan University of Technology,Zhengzhou Henan 450052,China) |
| |
| Abstract: | A new approach for symbolic rules extraction from support vector machines based on heuristic information was proposed,which solved the attribute selection and the division of attribute space.The position and shape characteristics of the classification hypersurface of Support Vector Regression(SVR)were used as heuristic information to direct the attribute selection and the division of attribute space.Then,the algorithm was given.Experiment results show that the proposed approach can improve the validity of t... |
| |
| Keywords: | heuristic information Support Vector Machine(SVM) Support Vector Regression(SVR) rule extraction |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机应用》浏览原始摘要信息 |
|
点击此处可从《计算机应用》下载全文 |
