Time Series Analysis Of Shanghai Copper Futures Price

With the development of China's market economy,the demand for value-added investment in wealth has also increased.China's financial trading market has thus prospered rapidly and there are more and more types of financial transactions.Managers of various institutions and investors as individuals have gradually focused on the emerging financial derivatives transactions,of which the futures market is particularly important to them.In the futures market,in order to avoid losses beyond the tolerable range,risk control is very important for everyone who trades in the futures market,and forecasting futures prices is a major prerequisite for risk control.Therefore,the importance of futures price forecasts is self-evident.Nowadays,with the continuous development of statistical theory and the continuous optimization of computer algorithms,many practical prediction methods have emerged for the futures price forecasting problem.On the one hand,there are statistical methods that extend to applications,such as ARIMA prediction models in time series,GM(1,1)gray prediction models,VAR models,and so on.On the other hand,there are methods related to machine learning algorithm theory,such as BP neural network prediction model and SVM support vector machine.At first,this paper selects the ARIMA prediction model in time series and the GM(1,1)gray prediction model.These two simple time series prediction models use the main force of Shanghai copper futures from August 3,2018 to December 14,2018.Cu1901)The price of the closing price of 90 consecutive trading days is modeled as raw data,and the Shanghai copper futures price of the same time period(December 17,2018 to December 28,2018)is predicted.The predicted values of the two predictive models are compared to the real values.According to the prediction model of this paper:the cumulative relative error value of ARIMA(8,2,1)model is smaller than that of GM(1,1)model,indicating that the prediction accuracy of ARIMA model is higher.This is mainly because the Shanghai copper futures price series does not meet the trend of increasing the index type,and is not suitable for constructing the gray forecasting model.In addition,this paper will construct an optimal weight linear combination forecasting model based on ARIMA(8,2,1)model and GM(1,1)model.Comparing the prediction accuracy of the optimal weight linear combination prediction model with the prediction accuracy of the ARIMA(8,2,1)model and the GM(1,1)model,it draws to a conclusion that the prediction accuracy of the combined prediction model is not better than all single prediction models' in the short-term prediction.