Research and Application of a Novel Nonlinear Grey Bernoulli Simpson Model for Short-Term Coke Production Forecasting

Abstract

Aims: As a basic energy source, coal occupies a leading position in the production and consumption of energy. If a reasonable coal energy production policy is to be formulated, effective forecasting is essential. Due to the lack of data, effective prediction with small samples has become the key to research. Study Design: A nonlinear grey Bernoulli Simpson model based on new information priority accumulation method is developed in this work to forecast the coke production in the Anhui China. The introduction of non-linear parameters makes the new model constructed with universality Place and Duration of Study: School of Science, Southwest University of Science and Technology, Mianyang, between April 2021 and June 2021. Methodology: This paper has established the nonlinear grey Bernoulli Simpson model with new information priority accumulation. Based on the grid search optimization, the data is divided by the leave-out method to construct a nonlinear problem to solve the nonlinear parameters of the model. Finally, the new model established was applied to the forecast of coke production in Anhui Province, China. Results: The MAPE and RMSPE of the nonlinear grey Bernoulli Simpson model based on new information priority accumulation method are 1.86% and 2.58%, which are lower than other comparative models. Conclusion: The application research of coke production shows that the new model proposed in this paper has the advantage of high prediction accuracy, which indicates that this method has great potential in the short-term prediction of energy production.