摘要
SVM是由Vapnik及其领导的AT&T Bell实验室研究小组提出的一种新的非常有发展前景的机器学习算法。本文通过它与神经网络学习算法的比较,说明了SVM具有较强的理论依据和较好的泛化性能。本文是SVM的发展综述,重点介绍了SVM的理论依据和一些值得研究的问题。
SVM is a new kind of promising machine learning algorithm proposed by Vapnik and his group at AT&T Bell laboratory. This paper is a survey of development in SVM, focusing on the theoretical foundation and some interesting problems of SVM. It also demonstrates that SVM has stronger theoretical foundation and better generalization performance by comparing it with the learning algorithms using neural networks.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2000年第3期285-290,共6页
Pattern Recognition and Artificial Intelligence
关键词
机器学习
神经网络
VC理论
SVM
学习算法
Machine Learning, Support Vector Machine, Neural Network, VC Theory, Empirical Risk, Expected Risk
