摘要
支持向量机(SVM)以其坚实的理论基础,和在机器学习领域表现出的良好推广性能,获得了越来越广泛的关注。为更好地推进其发展,科研工作者们借鉴统计学中经典的贝叶斯理论,做了大量工作,例如:引进贝叶斯理论中先验知识、后验概率等概念,改进支持向量机中的判别准则;或利用贝叶斯理论估计支持向量机中的参数w、正规化参数以及核参数等。目前已取得不错的效果,使支持向量机理论更具有实用价值。
Support Vector Machines（SVMs） are getting growing concerns due to its sound foundation of theories as well as its preferable popularising performance in the field of machine learning.In order to further promote its development,a lot of works have been doing by the scientific and technological personnel referring to classical Bayes＇ theorem in Statistics.For example,the concepts of priori knowledge and posterior probability in Bayes＇ theorem are introduced to improve the judging criterion on SVMs;or Bayes＇ theorem is employed to estimate the parameter w,normalisation parameter and kernel parameter of SVMs,etc.,and all of these have achieved quite satisfying effect,which makes the SVM theory more valuable in practice.In this paper,we are to summarise the works done in these areas.
出处
《计算机应用与软件》
CSCD
2010年第5期179-181,193,共4页
Computer Applications and Software
关键词
支持向量机
贝叶斯理论
先验概率
后验概率
Support vector machine Bayes＇ theorem Prior probability Posterior probability
