支持向量机的基本理论和研究进展  

The Basic Theory and Research Progress of Support Vector Machine

作　　者：林香亮 袁瑞[1] 孙玉秋[1] 王超[1] 陈长胜 

机构地区：[1]长江大学信息与数学学院,湖北荆州434023

出　　处：《长江大学学报（自然科学版）》2018年第17期48-53,共6页Journal of Yangtze University(Natural Science Edition)

基　　金：湖北省教育厅科学技术研究项目(Q20181310);油气资源与勘探技术教育部重点实验室(长江大学)开放基金项目

摘　　要：作为一种新的机器学习方法,依据结构风险最小原理,支持向量机表现出独特的泛化和推广能力,已逐渐成为国内外机器学习研究的热点之一。简要回顾了传统支持向量机的发展历史与基本理论,介绍了支持向量机的改进算法,系统总结了支持向量机在分类与回归问题中的具体应用实例及其优势。经过近30年的发展,出现了诸多改进的支持向量机算法,支持向量机的理论逐渐完善,其应用也得以深入各个研究领域,在解决小样本数据的分类与回归问题具有良好的应用优势,在智能故障诊断、图像处理、石油探勘与开发、说话人识别、水质检测与评价、金融预测、气象预测等方面获得了良好的应用效果。As a new kind of machine learning method,according to the principle of minimum structural risk,support vector machine(SVM)shows unique generalization and popularization ability.It has become one of the hotspots of machine learning research at home and abroad.The development history and basic theory of traditional SVM are briefly reviewed,the improved algorithm of support vector machine is introduced,the application examples and advantages of support vector machine(SVM)in classification and regression are summarized systematically.After nearly 30 years of development,a number of improved support vector machine algorithms have been emerged.The theory of support vector machines has been improved gradually,and its application has been introduced to various research fields.It has good application advantages for classification and regression of small sample data,good application effect is obtained in the intelligent trouble diagnosis,image processing,petroleum exploration and development,the speaker recognition,the water quality testing and evaluation,financial forecast,the weather forecast and so on.

关 键 词：支持向量机 分类 回归 应用 研究进展 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]