Improved financial time series forecasting by combining Support Vector Machines with self-organizing feature map

A two-stage neural network architecture constructed by combining Support Vector Machines (SVMs) with self-organizing feature map (SOM) is proposed for financial time series forecasting. In the first stage, SOM is used as a clustering algorithm to partition the whole input space into several disjoint regions. A tree-structured architecture is adopted in the partition to avoid the problem of predetermining the number of partitioned regions. Then, in the second stage, multiple SVMs, also called SVM experts, that best fit each partitioned region are constructed by finding the most appropriate kernel function and the optimal learning parameters of SVMs. The Santa Fe exchange rate and five real futures contracts are used in the experiment. It is shown that the proposed method achieves both significantly higher prediction performance and faster convergence speed in comparison with a single SVM model.

[1]  Andreas S. Weigend,et al.  Time Series Prediction: Forecasting the Future and Understanding the Past , 1994 .

[2]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[3]  Wei-Guo Cheng,et al.  "Forecasting the 30-year U.S. Treasury Bond with a System of Neural Networks" , 2000 .

[4]  Klaus-Robert Müller,et al.  Analysis of switching dynamics with competing neural networks , 1995 .

[5]  A. Weigend,et al.  Time Series Prediction: Forecasting the Future and Understanding the Past , 1994 .

[6]  Klaus-Robert Müller,et al.  Annealed Competition of Experts for a Segmentation and Classification of Switching Dynamics , 1996, Neural Computation.

[7]  Michael S. Schmidt,et al.  Identifying Speakers With Support Vector Networks , 1996 .

[8]  Thorsten Joachims,et al.  Text categorization with support vector machines , 1999 .

[9]  Ruy Luiz Milidiú,et al.  Time-series forecasting through wavelets transformation and a mixture of expert models , 1999, Neurocomputing.

[10]  Geoffrey E. Hinton,et al.  Adaptive Mixtures of Local Experts , 1991, Neural Computation.

[11]  S. Hyakin,et al.  Neural Networks: A Comprehensive Foundation , 1994 .

[12]  Alexander J. Smola,et al.  Support Vector Method for Function Approximation, Regression Estimation and Signal Processing , 1996, NIPS.

[13]  Sylvie Thiria,et al.  Sea surface temperature forecasts using on-line local learning algorithm in upwelling regions , 2000, Neurocomputing.

[14]  James T. Kwok.  Support vector mixture for classification and regression problems , 1998, ICPR.

[15]  Guido Deboeck,et al.  Trading on the Edge: Neural, Genetic, and Fuzzy Systems for Chaotic Financial Markets , 1994 .

[16]  Alexander J. Smola,et al.  Learning with kernels , 1998 .

[17]  Robert A. Jacobs,et al.  Hierarchical Mixtures of Experts and the EM Algorithm , 1993, Neural Computation.

[18]  Bernhard Schölkopf,et al.  Learning with kernels , 2001 .

[19]  Ramesh Sharda,et al.  Connectionist approach to time series prediction: an empirical test , 1992, J. Intell. Manuf..

[20]  Francis Eng Hock Tay,et al.  Financial Forecasting Using Support Vector Machines , 2001, Neural Computing & Applications.

[21]  Teuvo Kohonen,et al.  Self-organization and associative memory: 3rd edition , 1989 .

[22]  F. Girosi,et al.  Nonlinear prediction of chaotic time series using support vector machines , 1997, Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop.

[23]  Teuvo Kohonen,et al.  Self-Organization and Associative Memory , 1988 .

[24]  Francis Eng Hock Tay,et al.  A comparative study of saliency analysis and genetic algorithm for feature selection in support vector machines , 2001, Intell. Data Anal..