Predicting coal price using time series methods and combination of radial basis function (RBF) neural network with time series

This paper estimates the coal prices using two-time series and combined radial basis function (RBF) neural network methods. The time series method was simulated using the Monte Carlo simulation, while the combined RBF neural network method was employed using MATLAB software. The required data, including historical daily coal prices from 2018 to 2020, were collected. This hybrid method has eventually approximated the coal price with acceptable precision concerning the time series method. Due to the high differences among historical data in some periods, the root means squared error (RMSE), 1.49, and the correlation coefficient, 0.93651, of the combined model have provided a better prediction than the time series method RMSE, 2.68, and the correlation coefficient, 0.32. The results revealed that the combination of Brownian motion with mean return (BMMR) and RBF NN model (CBRN) has eventually been able to satisfactorily reduce the error value considering the data differences due to the critical factors, including economic and political conditions. The hybrid method can be used as an appropriate method for estimating prices in many financial markets.