An Optimizing Research For GEI Forecasting Based On Error Correction

As one of the most important indexes to reflect the overall price fluctuation of stock market, Stock price index, at the meantime, is one of the most important instruments in financial markets. So how to predict the price fluctuation of stock price index is always the focus question in the financial research all over the world.ARMA model and ARMA-GARCH model are two of the most commonly used models in the application of financial time series analysis. The hybrid ARMA-GARCH model achieves higher prediction accuracy than the sole ARMA model. Additionally, Support Vector Regression(SVR) model is used to fit the residuals of ARMA-GARCH model, and using GA and PSO algorisms to search the optimized parameters in SVR model. The result indicate that, both model precision of prediction optimized by GA and PSO have improvement compared with the one using default parameters, to some extent.