Journal of South China University of Technology (Natural Science Edition) ›› 2006, Vol. 34 ›› Issue (6): 64-68.

• Electronics, Communication & Automation Technology • Previous Articles     Next Articles

On-line M ulti-Output Support Vector Regression and Its Application to Investment Decision

Hu Gen-sheng  Deng Fei-qi   

  1. College of Automation Science and Engineering,South China Univ.of Teeh.,Guangzhou 510640,Guangdong,China
  • Received:2005-06-27 Online:2006-06-25 Published:2006-06-25
  • Contact: 胡根生(1971-)，男，博士生，主要从事支持向量机及其应用方面的研究 E-mail:hugs2906@sina.com
  • About author:胡根生(1971-)，男，博士生，主要从事支持向量机及其应用方面的研究
  • Supported by:

    国家自然科学基金资助项目(60374023)；广东省自然科学基金资助项目(011629)

PDF

770

Abstract:

An on-line multi output support vector machine regression algorithm is proposed in this paper.By using the gradient descent algorithm to minimize the instantaneous regularized risk of prediction results,the iterative for mulae of the weight coefficients and the bias of the regression function are obtained.Thus,the on-line multi-output regression prediction can be implemented for new arriving samples.The proposed algorithm is then applied to the investment decision to predict the optimal portfolio on line.Simulation results show that the proposed algorithm is easy to carry out because of its simple computation and small workload.

Key words: support vector regression, multi-output prediction, on-line learning, investment decision