| 四类基于支持向量机的多类分类器的性能比较 |
| |
| 引用本文： | 候明,张新新,范丽亚.四类基于支持向量机的多类分类器的性能比较[J].聊城大学学报(自然科学版),2014,27(3):54-60. |
| |
| 作者姓名： | 候明 张新新 范丽亚 |
| |
| 作者单位： | 聊城大学数学科学学院,山东聊城,252059 |
| |
| 基金项目： | 山东省高等学校科技计划项目 |
| |
| 摘 要： | 本文通过UCI数据库中的五个数据集,从分类准确性和运行时间两个方面,比较了四类基于支持向量机的多类分类器——多类支持向量机(MSVM),多元双生支持向量机(MTSVM),多生支持向量机(MBSVM)以及多元支持向量分类回归机(MSVCR)的性能.实验结果表明,MTSVM和MBSVM比MSVM和MSVCR更有效.
|
| 关 键 词： | 多类支持向量机 多元双生支持向量机 多生支持向量机 多元支持向量分类回归机 |
Comparison of Performance for Four Kinds of Multi-class Classifiers Based on Support Vector Machines |
| |
| HOU Ming,ZHANG Xin-xin,FAN Li-ya.Comparison of Performance for Four Kinds of Multi-class Classifiers Based on Support Vector Machines[J].Journal of Liaocheng University:Natural Science Edition,2014,27(3):54-60. |
| |
| Authors: | HOU Ming ZHANG Xin-xin FAN Li-ya |
| |
| Institution: | (School of Mathematical Sciences, Liaocheng University, Liaocheng 252059,China) |
| |
| Abstract: | In this paper,by means of five data sets in UCI database,performance of four kinds of multi-class classifiers,multiple support vector machine（MSVM）,multiple twin support vector machine（MTSVM）,multiple birth support vector machine（MBSVM）,multiple support vector classification regression（MSVCR）,is compared from classification accuracy and running time.The experiment results indicate that MTSVM and MBSVM are more effective than MSVM and MSVCR. |
| |
| Keywords: | multiple support vector machine multiple twin support vector machine multiple birth support vector machine multiple support vector classification reg ression |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
