| 一种结合K近邻法的改进的渐进直推式支持向量机学习算法 |
| |
| 引用本文： | 廖东平,王书宏,黎湘.一种结合K近邻法的改进的渐进直推式支持向量机学习算法[J].电光与控制,2010,17(10). |
| |
| 作者姓名： | 廖东平 王书宏 黎湘 |
| |
| 作者单位： | 国防科技大学电子科学与工程学院空间电子信息技术研究所,长沙,410073 |
| |
| 基金项目： | 国防预研基金资助课题 |
| |
| 摘 要： | 为了进一步提高改进的渐进直推式支持向量机学习算法(IPTSVML)的速度,提出了一种结合K近邻法(KNN)的改进的渐进直推式支持向量机学习算法,利用KNN对无标签样本集进行删减,去掉对学习作用不大的无标签样本,再对有标签样本集和剩余的无标签样本集利用IPTSVML算法进行学习与分类。雷达实测数据实验结果表明该算法是有效的。
|
| 关 键 词： | 统计学习理论 目标识别 直推式支持向量机(TSVM) K近邻法(KNN) |
An Improved Progressive TSVM Learning Algorithm Combined with K Nearest Neighbor |
| |
| LIAO Dongping,WANG Shuhong,LI Xiang.An Improved Progressive TSVM Learning Algorithm Combined with K Nearest Neighbor[J].Electronics Optics & Control,2010,17(10). |
| |
| Authors: | LIAO Dongping WANG Shuhong LI Xiang |
| |
| Affiliation: | LIAO Dongping,WANG Shuhong,LI Xiang(Institute of Space Electronics Information Technology,National University of Defense Technology,Changsha 410073,China) |
| |
| Abstract: | To speed up the Improved Progressive Transductive Support Vector Machine Learning Algorithm(IPTSVML),an IPTSVM combined with K Nearest Neighbor(KNN) is presented.First,KNN is used to prune the unlabeled example set for getting rid of the unlabeled examples that play a less important role in the learning process.Then the labeled example set and the other part of the unlabeled example set are learned and classified by using IPTSVML algorithm.Experiments on radar raw data showed the validity of this algorithm. |
| |
| Keywords: | statistical learning theory target recognition Transductive Support Vector Machine(TSVM) K Nearest Neighbor(KNN) |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
