| 模式分类方法研究 |
| |
| 引用本文： | 姜斌,黎湘,王宏强,郭桂蓉.模式分类方法研究[J].系统工程与电子技术,2007,29(1):99-102. |
| |
| 作者姓名： | 姜斌 黎湘 王宏强 郭桂蓉 |
| |
| 作者单位： | 1. 国防科技大学四院空间信息技术研究所,湖南,长沙,410073;中国人民解放军通信指挥学院,湖北,武汉,430010
2. 国防科技大学四院空间信息技术研究所,湖南,长沙,410073 |
| |
| 摘 要： | 分析了模糊K-近邻方法FKNN,研究了模糊最小最大神经网络FMMNN分类方法,然后深入分析了支持向量机SVM原理,并在此基础上给出了一种改进的径向基核函数。基于IRIS数据,进行了计算机仿真实验,结果表明,改进的SVM方法分类性能比模糊最小最大神经网络与模糊K-近邻算法的分类性能更好,且运算时间更短,更易于实时实现。
|
| 关 键 词： | 目标识别 模式分类 支持向量机 模糊K-近邻 神经网络 |
| 文章编号： | 1001-506X(2007)01-0099-04 |
| 修稿时间： | 2005年12月15 |
Methods of pattern classification |
| |
| JIANG Bin,LI Xiang,WANG Hong-qiang,GUO Gui-rong.Methods of pattern classification[J].System Engineering and Electronics,2007,29(1):99-102. |
| |
| Authors: | JIANG Bin LI Xiang WANG Hong-qiang GUO Gui-rong |
| |
| Abstract: | Pattern classification is not only an important aspect in pattern recognition but also the key technique in transaction of other questions.Firstly the method of FKNN is analyzed.Secondly the method of FMMNN is also studied.Thirdly studying the theory of SVM,this paper presents an ameliorated RBF kernel function.Finally,the simulation is processed based on the IRIS data and the results are compared with each other.These results show that the performance of the ameliorated SVM outperforms those of the FMMNN and FKNN,with shorter operation time and more suitable for real-time implementation. |
| |
| Keywords: | target recognition pattern classification support vector machine fuzzy K-nearest neightor neural network |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
