GARCH model prediction method based on Hessian matrix dynamic programming deep neural network.

One futures market GARCH model prediction method based on Hessian matrix dynamic programming deep neural networks has been proposed to improve the prediction accuracy of futures market model. Firstly, data analysis has been made for futures market based on GARCH model. It takes twelve factors of today opening, the highest price, the lowest price, today closing, ups and downs, volume, SMA, 10BIAS, 10PSY, 10RSI, 10AR and 10BR as input variables, which is comparative series and then set up GARCH prediction model; secondly, Hessian dynamic programming deep learning network has been constructed, learning and training have been made for the established futures market GARCH model to improve prediction accuracy and efficiency; lastly, the effectiveness of this algorithm has been verified through simulation experiment. [ABSTRACT FROM AUTHOR]