| 基于加权策略的SVM多元分类器 |
| |
| 引用本文： | 曹鸿,董守斌,张凌. 基于加权策略的SVM多元分类器[J]. 山东大学学报(理学版), 2006, 41(3): 66-69 |
| |
| 作者姓名： | 曹鸿 董守斌 张凌 |
| |
| 作者单位： | 华南理工大学,广东省计算机网络重点实验室,广东,广州,510640;华南理工大学,广东省计算机网络重点实验室,广东,广州,510640;华南理工大学,广东省计算机网络重点实验室,广东,广州,510640 |
| |
| 摘 要： | 多元分类器通常需要在训练时间和分类精度之间折衷.提出了加权阈值策略和一对多分类方法的改进算法 OVA WWT，以增加结果融合的公平性，从而提高分类精度.基于OVA WWT策略和SVMlight二元分类器，实现了基于SVMlight的多元分类器MSVMlight.在CWT100G数据集进行的实验表明，该分类器具有较高的分类精度以及较短的训练和分类时间.相同的数据集上的阈值策略选择实验也说明了加权阈值策略能提高分类精度.
|
| 关 键 词： | 支持向量机 一对多 加权阈值策略 多元分类 |
| 文章编号： | 1671-9352（2006）03-0029-04 |
| 收稿时间： | 2006-03-29 |
| 修稿时间： | 2006-03-29 |
A SVM multi-classifier based on the weighted threshold strategy |
| |
| CAO Hong,DONG Shou-bin,ZHANG Ling. A SVM multi-classifier based on the weighted threshold strategy[J]. Journal of Shandong University, 2006, 41(3): 66-69 |
| |
| Authors: | CAO Hong DONG Shou-bin ZHANG Ling |
| |
| Affiliation: | Guangdong Key laboratory of Computer Network, South China University of Technology,Cuangzhou 510640, Guangdong, China |
| |
| Abstract: | A weighted threshold strategy named WRCut and an improved OVA algorithm named OVA-WWT are presented to improvethe equitableness and the precision of classifiers, a multi-classifier of SVM^light named MSVM^light based on the OVA-WWT strategy is implemented, and two experiments on CWT100G data set are constructed, one to compare MSVM^light with other classitiers, and the other to compare WRCut strategy with RCut strategy. The results show that compared with other classifiers MSVM^light has a higher precision rate and shorter training time and that OVA-WWT algorithm can improve the precision rate of OVA. |
| |
| Keywords: | support vector machine one-vs-all weighted threshold strategy multi-classification |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《山东大学学报(理学版)》浏览原始摘要信息 |
|
点击此处可从《山东大学学报(理学版)》下载全文 |
|
