Analysis And Application Of Volatility Forecast Based On EMD Method

The financial market is constantly changing,China’s economy is constantly in line with international standards,the global economy is becoming more and more closely connected,and the volatility in the financial market is also increasing.Volatility is the embodiment of risk.For investors,it is only possible to estimate risk in order to make better investment decisions and asset portfolio choices in the market.With the continuous development of theory,there are many models and methods for measuring volatility,and the study of volatility in financial markets has also become an important topic.This paper proposes a new prediction method based on GARCH-MIDAS,that is,applying the empirical mode decomposition method to the prediction analysis of the stock market volatility,and combining the low frequency components and residual items obtained by the empirical mode decomposition method to represent fluctuating long-term trends.Another method is to decompose the original time series by the empirical modal method,and add different components to form a high-frequency component and a low-frequency component,and the residual term remains unchanged.Different methods are used to predict the high-frequency component,the low-frequency component,and the residual term,and the results obtained by adding the respective prediction results are added together to obtain the total predicted value.Finally,the method of model averaging is used to combine multiple models into a comprehensive model,and the prediction results of the models are compared and analyzed.At the same time,VaR is estimated in the application.The empirical results show that the IMF-GARCH model and the EMD method have higher prediction accuracy than the GARCH-MIDAS model,and the sequence 1 has the highest accuracy in the model.This point is also reflected in the estimation of VaR.The estimation of VaR based on the IMF-GARCH model and the EMD method improves the prediction accuracy.In summary,the empirical mode decomposition method is very effective for the prediction of financial time series.The non-linear characteristics of fused data make scholars continue to study new models and methods to make predictions,and reduce the noise in the original data through empirical decomposition methods,which is helpful for grasping the characteristics of financial volatility.This provides a new way of thinking and method for studying the fluctuation law in the financial market in the future.