Rebar Period Price Correlation And Research On Spot Price Forecast Based On Futures

With the ongoing impact of COVID-19 and international political and economic conflicts,steel prices have fluctuated sharply in recent years.It is necessary to avoid the risk of price fluctuation of rebar steel to ensure the healthy development of real economy.In financial theory,futures have the function of anticipatory price discovery,but the research on the numerical relationship between futures price and spot price and how to use futures price to make numerical prediction of spot price is still in its infancy.This paper selected in the past ten years,the main rebar contract in Shanghai Futures Exchange every trading day closing price,and the rebar spot market in Shanghai every trading day corresponding spot price as the original data,to explore the relationship between the current price of rebar period,and the forward price forecast of rebar in the spot market.First of all,we carry on the rebar spot price and futures price correlation research.The verification results show that the correlation coefficient between rebar spot price and futures price is 0.97,which has a strong positive correlation.Meanwhile,the P-value of Grange causality test is 2.3398E-07,which is far less than 0.05,indicating that there is a causal relationship between the two.It can be considered that futures price and spot price have a significant numerical correlation.Therefore,we consider the use of futures prices,rebar spot market future price forecast modeling.The prediction models based on ARIMA-GARCH model and BP neural network model were established respectively.On the basis of analyzing the advantages and disadvantages of each model,an improved model was proposed: the improved model based on BP neural network with the introduction of information entropy,and the model was evaluated by using the sum of squares of error and other indicators.Verification results show that the spot price prediction model based on ARIMA-GARCH model and BP neural network model,the relative errors of prediction results are kept in a small range,has a certain degree of desirability.However,the long-term prediction error of ARIMA-GARCH model is obviously large.In the long term,it can only make rough prediction of the price change trend with low accuracy.The prediction accuracy of BP neural network model is higher,and the stability of long-term prediction is better,but the prediction error is large in some time periods,and there is still room for improvement.The improved BP neural network model with the introduction of information entropy has the best prediction effect,which has high accuracy and good stability.The relative error of prediction is maintained at 1% for a long time,and the value of each model evaluation index is the minimum.The model has the best effect,which can achieve stable price prediction and has high practical application value.