Cross-Variety Arbitrage Strategy Of Ferrous Metal Futures Based On Genetic Algorithm Optimization Of Elman Neural Network

With the rapid development of China's economy and the continuous improvement of the relevant trading system,China's futures market has become increasingly mature and begins to move towards the goal of internationalization.More and more abundant futures varieties not only provide hedging tools for different industries,but also provide more trading products for arbitrageurs.This paper chooses the representative rebar,iron ore and coke futures in the ferrous metal industry chain as the research object of cross-variety arbitrage.Different from statistical arbitrage method,after choosing the appropriate arbitrage combination according to the results of cointegration test and using GARCH model to determine the arbitrage ratio,Elman neural network,which is better than BP neural network in dynamic system fitting,is used to predict the spread,and genetic algorithm is used to optimize it.Based on the prediction of the neural network,this paper constructs the trend strategy under different fixed thresholds.The relevant profit indicators show that with the increase of the opening threshold,the profitability of the neural network strategy decreases continuously,because the increase of the threshold makes the number of trades of the strategy decrease significantly.The strategy of Elman neural network based on Genetic algorithm optimization is better than that of BP neural network and Elman neural network.The network can produce higher annual rate of return,and the rate of return keeps stable growth in the long run.This is due to the fact that genetic algorithm accelerates the convergence speed of Elman neural network,increases the fitting accuracy of the network and makes the strategy generate more effective profit times.In addition,genetic algorithm enables the network to obtain the global optimal solution during training,so that the prediction can maintain a high accuracy in most time periods.Finally,this paper prospects the follow-up research of neural network and futures varieties,and provides specific investment suggestions for different market participants,which makes this paper not only have certain theoretical value,but also have strong practical significance.