Local Modeling Using Self-Organizing Maps and Single Layer Neural Networks
The paper presents a method for time series prediction using local dynamic modeling. After embedding the input data in a reconstruction space using a memory structure, a self-organizing map (SOM) derives a set of local models from these data. Afterwards, a set of single layer neural networks, trained optimally with a system of linear equations, is applied at the SOM’s output. The goal of the last network is to fit a local model from the winning neuron and a set of neighbours of the SOM map. Finally, the performance of the proposed method was validated using two chaotic time series.
KeywordsNormalise Mean Square Error Time Series Prediction Chaotic Time Series Desired Output Train Data
Unable to display preview. Download preview PDF.
- 2.Lawrence, S., Tsoi, A., Back, A.: Function approximation with neural networks and local methods: Bias, variance and smoothness. Australian Conference on Neural Networks (1996) 16–21Google Scholar
- 4.Takens, F.: Detecting strange attractors in turbulence. In Rand, D., Young, L., eds.: Dynamical Systems and Turbulence (Lecture Notes in Mathematics). Volume 898, New York, Springer-Verlag (1980) 365–381Google Scholar
- 6.Weigend, A.S., Gershenfeld, N.A.: Time Series Prediction: Forecasting the Future and Understanding the Past. Addison-Wesley, Reading, MA (1994)Google Scholar
- 7.Moshou, D., Ramon, H.: Extended self-organizing maps with local linear mappings for function approximation and system identification. Proceedings of the WSOM’97 (1997) 181–186Google Scholar
- 9.Kohonen, T.: Self-Organizing Feature Maps. Springer-Verlag, New York (1995)Google Scholar
- 10.Castillo, E., Fontenla-Romero, O., Alonso-Betanzos, A., Guijarro-Berdiñas, B.: A global optimum approach for one-layer neural networks. Neural Computation 14 (2002)Google Scholar