基于类中心的SVM 训练样本集缩减改进策略

doi:10.3969/j.issn.1674-0696.2014.02.35

重庆交通大学学报（自然科学版） ›› 2014, Vol. 33 ›› Issue (2): 154-158.DOI: 10.3969/j.issn.1674-0696.2014.02.35

• • 上一篇

基于类中心的SVM 训练样本集缩减改进策略

庞首颜^1，2，陈松¹，魏建猛¹，张元胜¹

1. 重庆交通大学信息科学与工程学院，重庆 400074; 2. 重庆第二师范学院，重庆 400065

收稿日期:2012-09-14 修回日期:2012-12-03 出版日期:2014-04-15 发布日期:2015-01-22
作者简介:庞首颜( 1987—) ，女，广西玉林人，硕士研究生，主要从事图形图像处理方面的研究。E-mail: PSY1239@126.com。

Reduction and Improvement Strategy of Training Sample Set for SVM Based on Class-Center

Pang Shouyan^1,2，Chen Song¹，Wei Jianmeng¹，Zhang Yuansheng¹

1. School of Information Science ＆ Engineering，Chongqing Jiaotong University，Chongqing 400074，China; 2. Chongqing University of Education，Chongqing 400065，China

Received:2012-09-14 Revised:2012-12-03 Online:2014-04-15 Published:2015-01-22

摘要/Abstract

摘要： 针对SVM 训练样本集规模较大引发的学习速度慢、存储需求量大、泛化能力降低等问题，通过改进的样本点到类中心的方法来确定边界样本，从而大量缩减训练样本，提高训练速度。此外，针对非线性空间无法直接通过计算得到特征空间类中心的问题，提出了一种通过在特征空间中，寻找能生成最小超球的样本点来近似代替特征样本的替代策略，使得在保证分类精度的同时，提高了训练速度。

关键词: 信息技术, SVM, 类中心, 边界样本, 替代策略

Abstract: SVM has practical applications in many fields for its superior performance at present． However，it has become a bottleneck to use SVM due to the problems including slow learning speed，large save requirement，and low generalization performance． Then an improvement reduction strategy was proposed based on the class-center，which significantly cut the amount of training samples and improved the training speed by determining the boundary training sample set． In addition，for the question that in the nonlinear space the center of the feature space class can not be directly obtained by calculating，an alternative strategy was proposed． In the alternative strategy，the sample spots that could generate the smallest hypersphere in the feature space which approximately substituted the center of feature samples were found． It improves the learning speed without reducing the classification accuracy．

Key words: information technology, Support Vector Machine( SVM) , class-center, boundary sample, alternative strategy

中图分类号:

TP391

庞首颜，陈松，魏建猛，张元胜. 基于类中心的SVM 训练样本集缩减改进策略[J]. 重庆交通大学学报（自然科学版）, 2014, 33(2): 154-158.

Pang Shouyan，Chen Song，Wei Jianmeng，Zhang Yuansheng. Reduction and Improvement Strategy of Training Sample Set for SVM Based on Class-Center[J]. Journal of Chongqing Jiaotong University(Natural Science), 2014, 33(2): 154-158.

参考文献

[1] 奉国和，李拥军，朱思铭．边界临近支持向量机[J].计算机应用研究，2006( 4) : 11-12．
Feng Guohe，Li Yongjun，Zhu Siming． Boundary nearest support vector machines [J].Application Ｒesearch of Computers，2006( 4) : 11-12．
[2] Cortes C，Vapnik V． Support-vector networks [J]. Machine Learning，1995， 20( 3) : 273-297.
[3] Osuna E，Freund Ｒ，Girosi F． An improved training algorithm for support vector machines[C] / / Principe J，Gile L，Morgan N， et al．Neural Networks for Signal Processing Ⅶ． Amelia Island，F L: Proceedings of the 1997 IEEE Workshop，1997: 276-285．
[4] Platt J C． Fast training of support vector machines using sequentia1 minima1 optimization[C] / /Schlkopf B，Burges C J C，Smola A J．Advances in Kernel Methods-Support Vector Learning． Cambridge: MIT Press， 1998: 185-208．
[5] Platt J C． Using analytic QP and sparseness to speed training of support vector machines [C] / /Kearns M S，Solla S A，Cohn D A． Advances in Neural Information Processing Systems． Cambridge: MIT Press， 1999: 557-563．
[6] 焦李成，张莉，周伟达．支撑矢量预选取的中心距离比值法[J].电子学报， 2001， 29( 3) : 383-386．
Jiao Licheng，Zhang Li，Zhou Weida． Pre-extracting support vectors for support vector machine [J].Chinese Journal of Electronics，2001， 29 ( 3) : 383-386．
[7] 汪西莉，焦李成．一种基于马氏距离的支持向量快速提取算法[J].西安电子科技大学学报: 自然科学版，2004，31 ( 4 ) : 639-643．
Wang Xili， Jiao Licheng． A fast algorithm for extracting the support vector on the mahalanobis distance [J]. Journal of Xidian University: Natural Science， 2004，31( 4) : 639-643．
[8] 李青，焦李成，周伟达．基于向量投影的支撑向量预选取[J].计算机学报，2005，28( 2) : 145-152．
Li Qing， Jiao Licheng，Zhou Weida． Pre-extracting support vector for support vector machine based on vector projection [J] ． Chinese Journal of Computers， 2005， 28( 2) : 145-152．
[9] 李红莲，王春花，袁保宗．一种改进的支持向量机NN_SVM[J].计算机学报， 2003， 26( 8) : 1015-1020．
Li Honglian，Wang Chunhua，Yuan Baozong． An improved SVM: NN_SVM [J].Chinese Journal of Computers，2003，26 ( 8 ) : 1015-1020．
[10] 郭亚琴，王正群．一种改进的支持向量机BS_SVM[J].微电子学与计算机， 2010， 27( 6) : 54-56．
Guo Yaqin，Wang Zhengqun． An improved SVM: BS _SVM [J].Microelectronics ＆ Computer，2010， 27( 6) : 54-56．
[11] 朱方，顾军华，杨欣伟，等．一种新的支持向量机大规模训练样本集缩减策略[J] ．计算机应用，2009，29( 10) : 2736-2740．
Zhu Fang，Gu Junhua，Yang Xinwei， et al． New reduction strategy of large-scale training sample set for SVM [J].Journal of Computer Applications， 2009， 29( 10) : 2736-2740．
[12] 罗瑜．支持向量机在机器学习中的应用研究[D] ．成都: 西南交通大学，2007．
Luo Yu． Study on Application of Machine Learning Based on Support Vector Machine[D] ． Chengdu: Southwest Jiaotong University，2007．
[13] 曹淑娟，刘小茂，张钧，等．基于类中心思想的去边缘模糊支持向量机[J] ．计算机工程与应用， 2006， 42( 22) : 146-149．
Cao Shujuan，Liu Xiaomao，Zhang Jun，et al． Fuzzy support vector machine of dismissing margin based on the method of class-center[J] ． Computer Engineering and Applications， 2006，42( 22) : 146-149．

基于类中心的SVM 训练样本集缩减改进策略

Reduction and Improvement Strategy of Training Sample Set for SVM Based on Class-Center

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 4

编辑推荐

Metrics

[1]	高金贺1，黄伟玲2，蒋浩鹏3. 城市交通碳排放预测的多模型对比分析[J]. 重庆交通大学学报（自然科学版）, 2020, 39(07): 33-39.
[2]	欧居尚. 道路交通管理信息化建设存在的问题及对策分析[J]. 重庆交通大学学报（自然科学版）, 2006, 25(2): 96-99.
[3]	罗川. 论现代知识管理与信息技术[J]. 重庆交通大学学报（自然科学版）, 2005, 24(4): 145-148.
[4]	董千里. 货运集散业及支持系统探讨[J]. 重庆交通大学学报（自然科学版）, 1995, 14(3): 47-52.