计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (32): 53-56.DOI: 10.3778/j.issn.1002-8331.2008.32.016

• 理论研究 • 上一篇    下一篇

支持向量回归机模型结构及性能的研究

司利云,林 辉   

  1. 西北工业大学 自动化学院,西安 710072
  • 收稿日期:2007-12-12 修回日期:2008-03-26 出版日期:2008-11-11 发布日期:2008-11-11
  • 通讯作者: 司利云

Study on model structure and approximation performance of Support Vector Regression

SI Li-yun,LIN Hui   

  1. Automatic College of Northwestern Polytechnical University,Xi’an 710072,China
  • Received:2007-12-12 Revised:2008-03-26 Online:2008-11-11 Published:2008-11-11
  • Contact: SI Li-yun

摘要: 总结了目前提出的四类不同结构的支持向量回归机(Support Vector Regression,SVR)模型,并在此基础上提出了三种新的模型结构,通过试验手段对比了几种典型模型的逼近性能,分析了各个模型在应用中存在的优缺点,为实际应用中采用合适的模型提供一种选择方法。

关键词: 支持向量回归机, 模型结构, 函数逼近性能

Abstract: The four kinds of model structures of Support Vector Regression(SVR) are introduced,and the three new model structures are presented.The approximation performance of several typical SVR functions is discussed by trial.This paper analyses relative merits of these typical functions,therefore provides one choice method for optimal SVR according to actual condition.

Key words: Support Vector Regression(SVR), model structure, approximation performance