Quantitative Investment Strategy Of CSI 300 Stock Index Futures Based On NARX Dynamic Neural Network

The intelligent robot AlphaGo developed by DeepMind laboratory defeated the Go world champion Lee Sedol in 2016.Artificial intelligence has set off a wave of AI in the field of computer science.Computer science and technology are going through a revolution.Simmons' medallion fund has made a splash in the private equity world with its quantitative strategies.The concept of quantitative investing is starting to spread through the investment community.Financial investment technology is also undergoing a revolution.If the current hot AI technology is used to study quantitative strategy,it will undoubtedly be a revolutionary promotion for the interdisciplinary study of computer technology and financial technology.Quantitative strategy model can deeply analyze the time-series data on the market.As a branch of artificial intelligence,neural network technology has excellent ability to deal with the problems of nonlinear systems.As we all know,the securities market is a complex and nonlinear system with noise.In recent years,the application of neural network technology in quantitative investment strategy has shown good results.This paper first compares the accuracy of the traditional static BP neural network and NARX dynamic neural network in predicting the opening price time series of CSI 300 stock index futures.Through the empirical analysis of the opening price time series of CSI 300 stock index futures,it is proved that the prediction fitting degree and accuracy of the NARX dynamic neural network are better than the traditional static BP neural network in the processing of time series of long time period data.Finally,based on the NARX dynamic neural network,this paper constructs a quantitative investment strategy to find the price difference between the closing price of the day and the opening price of the next trading day for the CSI 300 stock index futures.The results obtained by the NARX dynamic neural network prediction model show that the return rate of the quantitative strategy is better than the market's return rate,and the application of the dynamic neural network to the prediction of the time series of the CSI 300 stock index futures data is feasible and effective.