a fuzzy support vector machine algorithm with dual membership based on hypersphere
文献类型：期刊论文
| 作者 | Ding Shifei ; Gu Yaxiang |
| 刊名 | Journal of Computational Information Systems
 |
| 出版日期 | 2011 |
| 卷号 | 7期号:6页码:2028-2034 |
| 关键词 | Algorithms Statistics Support vector machines Vectors |
| ISSN号 | 1553-9105 |
| 中文摘要 | In traditional fuzzy support vector machine(FSVM), membership function is established in global scope will reduce the membership of support vectors, and the FSVM based dismissing margin increases the training speed, but will remove some support vector artificially. So, a new algorithm of Fuzzy Support Vector Machine with Dual Membership based on Hypersphere (HDM-FSVM) is proposed. In this algorithm, the two classes of hyperspheres are divided into two parts respectively. Then, according to most support vectors are in the hemispheres which close together, we use the membership function that can enhance the membership of support vector, and because of there are a few of support vectors in other hemispheres, we must ensure the high membership of support vectors and reduce the membership of non-support vector. In order to removal noise and outliers, we introduce a radius controlling factor to control size of hyperspheres, the samples that outside of hyperspheres are considered as noise and outliers. Experimental results show that HDM-FSVM can enhance the classification accuracy rate of the sample sets that contain noise and outliers. Copyright © 2011 Binary Information Press. |
| 英文摘要 | In traditional fuzzy support vector machine(FSVM), membership function is established in global scope will reduce the membership of support vectors, and the FSVM based dismissing margin increases the training speed, but will remove some support vector artificially. So, a new algorithm of Fuzzy Support Vector Machine with Dual Membership based on Hypersphere (HDM-FSVM) is proposed. In this algorithm, the two classes of hyperspheres are divided into two parts respectively. Then, according to most support vectors are in the hemispheres which close together, we use the membership function that can enhance the membership of support vector, and because of there are a few of support vectors in other hemispheres, we must ensure the high membership of support vectors and reduce the membership of non-support vector. In order to removal noise and outliers, we introduce a radius controlling factor to control size of hyperspheres, the samples that outside of hyperspheres are considered as noise and outliers. Experimental results show that HDM-FSVM can enhance the classification accuracy rate of the sample sets that contain noise and outliers. Copyright © 2011 Binary Information Press. |
| 收录类别 | EI |
| 语种 | 英语 |
| 公开日期 | 2013-10-08 |
| 源URL | [http://ir.iscas.ac.cn/handle/311060/16171]  |
| 专题 | 软件研究所_软件所图书馆_期刊论文 |
| 推荐引用方式 GB/T 7714 | Ding Shifei,Gu Yaxiang. a fuzzy support vector machine algorithm with dual membership based on hypersphere[J]. Journal of Computational Information Systems,2011,7(6):2028-2034. |
| APA | Ding Shifei,&Gu Yaxiang.(2011).a fuzzy support vector machine algorithm with dual membership based on hypersphere.Journal of Computational Information Systems,7(6),2028-2034. |
| MLA | Ding Shifei,et al."a fuzzy support vector machine algorithm with dual membership based on hypersphere".Journal of Computational Information Systems 7.6(2011):2028-2034. |
入库方式： OAI收割
来源：软件研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。