GARCH Estimated by Evolutionary Programming and Its Application on Stock Return Volatility

This paper constructs a GARCH regression model estimated by evolutionary programming (EP) for modeling the stock return volatility. GARCH has a strong ability to capture some of the typical stylized facts of financial time series, for example volatility clustering, which describes the tendency for volatility periods with similar magnitude to cluster. On the other hand, because the traditional estimation methods are complex and have many other shortcomings such as difficulty of selecting the starting values while EP can be implemented with ease and has a powerful optimizing performance, EP is employed to optimize the coefficients of GARCH regression model. Moreover, we evaluate the ability to forecast stock return volatility using Shanghai Stock Price Index and the experiment results reveal that our proposed model can efficiently capture the volatility effects.