This paper proposes a novel model of support function machine (SFM) for time series predictions. Two machine learning models, namely, support vector machines (SVM) and procedural neural networks (PNN) are compared in solving time series and they inspire the creation of SFM. SFM aims to extend the support vectors to spatiotemporal domain, in which each component of vectors is a function with respect to time. In the view of the function, SFM transfers a vector function of time to a static vector. Similar to the SVM training procedure, the corresponding learning algorithm for SFM is presented, which is equivalent to solving a quadratic programming. Moreover, two practical examples are investigated and the experimental results illustrate the feasibility of SFM in modeling time series predictions.


support vector machine learning algorithm support function procedure neural networks time series predictions 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Vapnik, V.: The nature of statistical learning theory. Springer, New York (1995)zbMATHGoogle Scholar
  2. 2.
    Scholkopf, B., Sung, K.K., Burges, C.J.C., Girosi, F., Niyogi, P., Poggio, T., Vapnik, V.: Comparing support vector machines with Gaussian kernels to radial basis function classifiers. IEEE Trans. Signal Process. 45, 2758–2765 (1997)CrossRefGoogle Scholar
  3. 3.
    Yang, Z.R., Chou, K.-C.: Bio-support vector machines for computational proteomics. Bioinformatics 20, 735–741 (2004)CrossRefGoogle Scholar
  4. 4.
    Vapnik, V.: The support vector method of function estimation. In: Suykens, J.A.K., Vandewalle, J. (eds.) Nonlinear Modeling: Advanced Black-Box Techniques, pp. 55–85. Kluwer, Boston, MA (1998)Google Scholar
  5. 5.
    Liang, J.Z, Han, J.M.: Complex number procedure neural networks. In: Wang, L., Chen, K., Ong, Y.S. (eds.) ICNC 2005. LNCS, vol. 3610, pp. 336–339. Springer, Heidelberg (2005)Google Scholar
  6. 6.
    Liang, J.Z.: Functional procedure neural networks. Dynamic of Continuous Discrete and Impulsive Systems-Series B-Applications & Algorithms 1 (S.I.), pp. 27–31 (2005)Google Scholar
  7. 7.
    Liang, J.Z., Zhou, J.Q., He, X.G.: Procedure neural networks with supervised learning. In: The 9th International Conference on Neural Information Processing, Singapore, pp. 523–527 (2002)Google Scholar
  8. 8.
    He, X.G., Liang, J.Z.: Some theoretic problems of procedure neural networks. Engineering Science in China 2, 40–44 (2000)Google Scholar
  9. 9.
    He, X.G., Liang, J.Z., Xu, S.H.: Training and application of procedure neural networks. Engineering Science in China 3, 31–35 (2001)Google Scholar
  10. 10.
    Flake, G., Lawrence, S.: Efficient SVM regression training with SMO. Machine Learning 46, 271–290 (2002)zbMATHCrossRefGoogle Scholar
  11. 11.
    Nello, C., John, S.T.: An introduction to support vector machines and other kernel-based learning methods. Cambridge University Press, Cambridge (2000)Google Scholar
  12. 12.
    Lawrence, R.: Using neural networks to forecast stock market prices. Reports (1997),
  13. 13.
  14. 14.
    Chan, A., Vasconcelos, N., Moreno, P.J.: A family of probabilistic kernels based on information divergence. Technical Report SVCL-TR-2004-01 (June 2004)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Jiuzhen Liang
    • 1
  1. 1.School of Information Technology, Jiangnan University, 1800 Lihu Road, Wuxi, Jiangsu Province, CHINA 214122 

Personalised recommendations