Coincidence of the solutions of the modified problem with the original problem of v-MC-SVM

  • Authors:

  • Xin Xue;Taian Liu;Xianming Kong;Wei Zhang

  • Affiliations:

  • College of Mathematics and System Science, Taishan University, Tai'an, P.R.China;Department of Information and Engineering, Shandong University of Science and Technology, Taian, P.R.China;College of Mathematics and System Science, Taishan University, Tai'an, P.R.China;College of Mathematics and System Science, Taishan University, Tai'an, P.R.China

  • Venue:
  • ISNN'10 Proceedings of the 7th international conference on Advances in Neural Networks - Volume Part II
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Multi-class support vector machine with parameter v (v-MC-SVM)is a kind of machine learning method which is similar to QP-MC-SVM The constraints of v-MC-SVM and it's dual problem are too complicated By adding the term bm to the objective function of v-MC-SVM, the original problem was modified Then,by employing the Kesler's construction, the modified problem is simplified Efficient algorithms can be used to solve the simplified problem Numerical testing results show that v-MC-SVM has the same accuracy rate as QP-MC-SVM Basing on Lagrange function and KKT condition, this paper proves strictly that the solution of the modified problem is the solution of the original problem, which provides the muli-class SVM with theory bases.