Application of Mixture of Experts Model To Financial Time Series Forecasting

Presently, there are many models to predict the trend of a time series. However, different models have different prediction ability. Combinations of these models may provide a better performance than those provided by individuals. In this paper, experts are constructed by choosing different architectural parameters of two modified network models: Buffered Back-Propagation and Improved Clusnet. Outputs of these experts are combined by Mixture of Experts model (Xu, Jordan and Hinton, 1994).