Sparrow-based optimised XGBoost blast furnace utilisation factor forecasting model

Abstract

An XGBoost regression blast furnace utilisation coefficient forecasting model based on sparrow optimisation is constructed. The model is to improve the blast furnace utilisation coefficient. To combine the advantages of sparrow-optimised XGBoost in terms of multiple linearity and regression, etc., and based on 652 sets of original sample information, we regress the blast furnace utilisation coefficient forecasting model with the help of sparrow-optimised XGBoost. The relationship between the regulation of economic and technical indicators (coke ratio, coal ratio, fuel ratio, etc.) and the blast furnace utilisation factor is thus determined. The accuracy of the blast furnace utilisation coefficient prediction is 93.1298%, and the RMSE of the sparrow-optimised XGBoost blast furnace utilisation coefficient prediction model is 0.2860, R2 is 0.7637 and MAPE is 0.0808. The simulation effect is effective and has considerable application prospect and promotion value in steel enterprises.