Fractal Characteristics And Price Prediction Of China's Rebar Market

With the rapid development of China's economy,the demand for steel is expanding,making the steel market gradually occupy an important position in the real economy.The ups and downs of steel prices may cause heavy losses to the real economy at any time.It is especially important to explore the internal structure of China's steel market and make price forecasts.Due to the wide variety of steel products in the steel market,this paper selects the most representative rebar as the research object.Considering the nonlinear characteristics of the rebar market,this paper will introduce the fractal theory and wavelet denoising method to explore the futures and spot markets of rebar in China,and predict the price of the futures and spot markets of rebar according to the fractal characteristics.Firstly,after using wavelet denoising to eliminate the influence of noise,this paper studied the futures and spot markets of rebar respectively by introducing modified R/S analysis method and OSW-MF-DFA method,and found that both the futures and spot markets of rebar had positive sustainability,and the strength of positive sustainability of spot was greater.In addition,after wavelet denoising,the strength of fractal characteristics of both rebar futures and spot become stronger.Secondly,this paper introduces the overlapping smooth window technique into the asymmetric multifractal detrended cross-correlation analysis(MF-ADCCA),and combines the multiscale multifractal analysis method(MMA)to propose an asymmetric multiscale multifractal detrended cross-correlation analysis method(A-MMDCCA).This not only makes up for the defect of pseudo-fluctuation in traditional methods,but also can better explore the asymmetric multifractal cross-correlation between the futures and spot markets of rebar in different scales.The results show that there is an asymmetric multifractal cross-correlation between rebar futures and the spot market.When the scale value is small,the cross-correlation between rebar futures and the spot market is the most complicated;when the rebar futures market is on the rise,its cross-correlation with the spot market is more significant.Finally,on the basis of wavelet denoising,combined with the fractal characteristics of the data,BP neural network model based on particle swarm optimization algorithm is constructed to predict the price of rebar futures and spot respectively.By comparing the prediction errors of the model,it can be found that the accuracy of the prediction model combined with wavelet de-noising is improved,which can better fit the price trend of the futures and spot markets of rebar,and the prediction accuracy of the spot market of rebar is higher.This paper describes the characteristic information of China's rebar futures and spot markets by non-linear method,which not only provides the reference of price rules for China's rebar enterprises and investors,but also provides the theoretical basis for market regulators to make policy.It is helpful to adjust the fluctuation range of the price of rebar so as to reduce the economic risk caused by the price fluctuation of rebar.