[1]赵春晖,陈万海,郭春燕.多类支持向量机方法的研究现状与分析[J].智能系统学报,2007,2(2):11-17.
　ZHAO Chun-hui,CHEN Wan-hai,GUO Chun-yan.Research and analysis of methods for multiclass support vector machines[J].CAAI Transactions on Intelligent Systems,2007,2(2):11-17.
点击复制

多类支持向量机方法的研究现状与分析

《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷: 2 期数: 2007年第2期 页码: 11-17 栏目: 综述 出版日期: 2007-04-25
Title:
Research and analysis of methods for multiclass support vector machines
文章编号:
1673-4785(2007)02-0011-07
作者:
赵春晖,陈万海,郭春燕
哈尔滨工程大学信息与通信工程学院，黑龙江哈尔滨150001
Author(s):
ZHAO Chun-hui, CHEN Wan-hai, GUO Chun-yan
College of Information and Communication Engineering, Harbin Engineering Univer sity, Harbin 150001, China
关键词:
多类支持向量机; 两类分类器; 层次结构; 一次性优化; 纠错编码
Keywords:
multiclass support vector machine;  binary classifier;  hierarchical str ucture;  oneoff optimization;  error correcting code
分类号:
TP391
文献标志码:
A
摘要:
支持向量机（SVM）是建立在统计学理论基础上的一种小样本机器学习方法，最初应用于解决两类分类问题.然而在解决实际问题中遇到的多为多分类问题，如何有效的将其推广到多类分类问题是一个正在研究的问题.该文对现有的多类支持向量机方法从组合多个两类分类器、层次结构、一次性优化问题和纠错编码等4个角度进行了综合归纳和分析，详细介绍了每种方法的代表性算法，并比较其优劣.
Abstract:
The SVM is a limited sample learning method which was developed from s tatistical theory, and originally designed for binary classification. However, m any practical problems are multiclassification ones. How to effectively extend binary classification to multiclassification is an ongoing research issue. Thi s paper generalizes and analyzes multiclass support vector machines from four ang les: combination of several binary classifiers, hierarchical structures, one-off optimization and error correcting codes. Several representative algorithms for various methods are introduced in detail and their advantages and disadvantages are compared.
参考文献/References:
［1］ HSU C W，LIN C J. A comparison of methods for multiclass support vector m achines［J］. IEEE Transactions on Neural Networks, 2002, 13(2): 415-425.
［2］KIJSIRKUL B, USSIVAKUL N. Multiclass support vector machines using adaptiv e directed acyclic graph［A］. Proceedings of the 2002 International Joint Confere nce on Neural Networks［C］.Honolulu, HI, USA, 2002,1(5): 980-985.
［3］PHETKAEW T, KIJSIRIKUL B, RIVEPIBOON W. Reordering adaptive directed acycl ic graphs: an improved algorithm for multiclass support vector machines［A］. Proc ee dings of the 2003 International Joint Conference on Neural Networks［C］.Portlan d, OR, USA, 2003.
［4］ TIAN X,DENG F Q. An improved multiclass SVM algorithm and its appli ca tion to the credit scoring model［A］. Proceedings of the fifth World Congress o n Intelligent Control and Automation［C］.Hangzhou, China,2004.
［5］ANGUITA D, RIDELLA S, STERPI D. A new method for multiclass support vecto r machines［A］. Proceedings of the 2004 International Joint Conference on Neural Networks［C］.Budapest, Hungary, 2004.
［6］黄? 勇,郑春颖,宋忠虎.多类支持向量机算法综述［J］. 计算技术与自动化, 2005, 24(4): 61-63.
?HUANG Yong, ZENG Chunyin, SONG Zhonghu. Multiclass support vector machines alg orithm summarization［J］. Computing Technology and Automation, 2005, 24(4): 6 1-63.
［7］刘志刚, 李德仁, 秦前清, 史文中. 支持向量机在多类分类问题中的推广［J ］. 计算机工程与应用, 2004(7): 10-13.
?LIN Zhigang, LI Deren, QIN Qianqing, SHI Wenzhong. An analytical overview of met hods for multicategory support vector machines［J］. Computer Engineering and A pplications, 2004(7): 10-13.
［8］TAKAHASHI F. Decisiontreebased multiclass support ve ctor mac hines［A］. Pr oceedings of the 9th International Conference on Neural Information Processing［ C］.Orchid Country Club, Singapore, 2002.
［9］PLATT J, CRISTIANINI N, SHAWETAYLOR J. Large margin DAGs for multiclass c la ssification［A］. Proceedings of Neural Information Processing Systems［C］.［s. l.］, Cambridge, 2000.
［10］PAUL G.Cox,Reza Adhami. Multiclass support vector machine class ifier app li ed to hyperspectral data［A］. Proceedings of the ThirtyFourth Southeastern Symposium on System Theory［C］.Huntsville, Alabama, 2002.
［11］FRANC V,HLAVAC V. Multiclass support vector machine ［A］. Proceedings o f the 16th International Conference on Pattern Recognition［C］.Quebec City, Can ada,2002. 
［12］LIN Chunfu,WAMG Shengde. Fuzzy support vector machines［J］. IEEE Transac tions on Neural Networks, 2002, 13(2):464-471.
［13］〖JP3〗LI Kunlun,HUANG Houkuan,TIAN Shengfeng. A novel multiclass SVM cl a ssifie r based on DDAG［A］. Proceedings of the 2002 International Conference on Machine Learning and Cybernetics［C］.Beijing, China,2002.
［14］ARENASGARCIA J,PEREZCRUZ F.Multiclass support vector machines: a new app roach［A］. Proceedings of the 2003 IEEE International Conference on Acoustics, Speech and Signal Processing［C］.Hong Kong,China,2003.
［15］LIU Yi,ZHENG Y F. Oneagainstall multiClass SVM classification using reli ability measures［A］. Proceedings of the 2005 International Joint Conference on Neural Networks［C］. Montreal, Canada,2005.
［16］LU Baoliang,WANG Kaian, UTIYAMA M. A partversuspart method for massive ly parallel training of support vector machines［A］. Proceedings of the 2004 Inte rnational Joint Conference on Neural Networks［C］. Budapest, Hungary,2004.
［17］XIN Dong, WU Zhaohui, PAN Yunhe. A new multiclass support vector machin es ［A］. IEEE International Conference on Systems, Man and Cybernetics［C］. Tucso n, USA,2001 .
［18］童学锋, 石繁槐. FSVM在有限集脱机手写体汉字识别中的应用［J］. 计算机工程, 2003, 29(3): 109-111.
TONG Xuefeng, SHI Fanhuai. Application 0f FSVM for limitedset offline handw rit ten chinese characters recognition［J］. Computer Engineering, 2003, 29(3): 109- 111.
［19］李昆仑, 黄厚宽, 田盛丰. 一种基于有向无环图的多类SVM分类器［J］. 模式识别与人工智能, 2003, 16(2): 164-168.
?LI Kunlun，HUANG Houkuan，TIAN Shengfeng. A novel multiclass SVM classifier ba s ed on DDAG［J］. Pattern Recognition and Artificial Intelligence, 2003, 16(2): 164-168.
［20］安金龙, 王正欧, 马振平. 一种新的支持向量机多类分类方法［J］. 信息与控制, 2004, 33(3): 262-267.
AN Jin1ong, WANG Zhengou, MA Zhenping. A new SVM multiclass classification metho d［J］. Information and Control, 2004, 33(3): 262-267.
［21］唐发明, 王仲东, 陈绵云. 支持向量机多类分类算法研究［J］. 控制与决策 , 2005, 20(7): 746-749.
TANG Faming, WANG Zhongdong, CHEN Mianyun. On multiclass classification m ethods for support vector machines［J］. Control and Decision, 2005, 20(7): 746-749.
［22］郑勇涛, 刘玉树. 支持向量机多分类问题研究［J］. 计算机工程与应用, 20 05, 41(23): 190-192.
?ZHENG Yongtao, LIU Yushu. An analysis of multiclass support vector machines［J ］. Computer Engineering and Applications, 2005, 41(23): 190-192.
［23］祁亨年. 支持向量机及其应用研究综述［J］. 计算机工程, 2004, 30(10): 6-9.
QI Hengnian. Support vector machines and application research overview［J］. Com puter Engineering, 2005, 41(23): 190-192.
［24］宋晓宁, 束 鑫. 一种改进的模糊支持向量机的人脸识别方法［J］. 微机发展, 2005, 15(3): 23-25.
SONG Xiaoning, SHU Xin. An improved fuzzy support vector machine and its applica tion to face recognition［J］. Microcomputer Development, 2005, 15(3): 23-25.
［25］刘江华, 程君实, 陈佳品. 支持向量机训练算法综述［J］. 信息与控制, 20 02, 31(1): 45-50.
?LIU JIanghua, CHENG Junshi, CHEN Jiapin. Support vector machine traning algorith m: a review［J］. Information and Control, 2002, 31(1): 45-50.
［26］张学工. 关于统计学习理论与支持向量机［J］. 自动化学报, 2000, 26(1) : 32-42.
ZHANG Xuegong. Introduction to statistical learning theory and support vector ma chines［J］. Acta Automatica Sinica, 2000, 26(1): 32-42.

备注/Memo

收稿日期：2006-11-14.
基金项目：
国家自然科学基金资助项目(60672034);高校博士点基金资助项目(2006 0217021);
黑龙江省自然科学基金资助项目(ZJG20606-01).
作者简介：
赵春晖,男，1965年生，博士，教授，博士生导师，主要研究方向为智能信息处理技术、图像处理.获省部级科技进步奖5项，发表论文180多篇，出版著作3部.E-mail:zhaochunhui@hrbeu.edu.cn.
陈万海，男，1963年生，副教授，信号与信息处理专业博士研究生，主要研究方向为超光谱遥感图像处理技术,发表论文8篇.
?郭春燕，女，1984年生，硕士研究生，主要研究方向为超光谱遥感图像处理技术、目标分类.

更新日期/Last Update: 2009-05-06
Copyright © 《 智能系统学报》 编辑部
地址：（150001）黑龙江省哈尔滨市南岗区南通大街145-1号楼 电话：0451- 82534001、82518134　邮箱：tis@vip.sina.com