Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (27): 216-218.DOI: 10.3778/j.issn.1002-8331.2009.27.066

• 工程与应用 • Previous Articles     Next Articles

Novel construction method for FLANN based on Support Vector Machina and its application

MAO Xian-bai,WANG Li-heng,LI Chang-xi   

  1. Department of Control Science and Engineering,Huazhong University of Science and Technology,Wuhan 430074,China
  • Received:2008-05-22 Revised:2008-08-22 Online:2009-09-21 Published:2009-09-21
  • Contact: MAO Xian-bai

基于SVM回归的FLANN改进新法及其应用

毛先柏,王利恒,李昌禧   

  1. 华中科技大学 控制科学与工程系,武汉 430074
  • 通讯作者: 毛先柏

Abstract: The equivalence of the generic Functional Link Artificial Neural Network(FLANN) and the Support Vector Machine(SVM) is given by analyzing their relations.A novel construction method for FLANN based on SVM is discussed and applied to dynamical compensation for weighting sensor.The results show that this method has the characteristics of unique result,simple structure and global optimum.Using this method to compensate weighting sensor dynamically can get obvious effect.It has the practical value in practice.

Key words: Functional Link Artificial Neural Network(FLANN), Support Vector Machine(SVM), dynamical compensation

摘要: 通过分析常规函数链接型神经网络(FLANN)结构与支持向量机(SVM)的关系,确定了两者本质上的等价性;在此基础上提出了一种基于SVM技术的FLANN构造新方法,并将SVM-FLANN应用到称重传感器的动态补偿上。结果表明该方法构造的FLANN具有结果唯一、结构简单、全局优化等特点,应用于称重传感器的动态补偿时,对传感器的性能改善效果明显,具有实用价值。

关键词: 函数链接型神经网络, 支持向量机, 动态补偿

CLC Number: