Recently, financial institutions and investors have placed an increasing emphasis on ESG (environmental, social, and governance) as a principal indicator for the evaluation of companies. However, the current ESG scoring systems lack uniformity and are often subjective. It is of great importance to be able to make accurate predictions regarding the ESG scores of corporations. A Stacked Generalization Model that employs Random Forest (RF), Gradient Boosting Decision Tree (GBDT), eXtreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM) as base learners, with Bayesian Ridge Regression (BRR) as the meta-model for integrating the predictions of these diverse models is proposed. The goal is to develop an ESG score prediction model for Chinese companies. The experimental data set encompasses Chinese A-share listed companies from 2012 to 2020. The Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (R²) are employed for model evaluation and are compared with seven benchmark models. The results demonstrate that SGM-BRR reduces the RMSE by 18.4%, 17.3%, 13.7%, and 76.1%, the MAE by 15.4%, 18.4%, 15.8%, and 68.4%, and increases the R² by 2%, 1.4%, 2%, and 6% for ESG, E, S, and G scores, respectively. Furthermore, the model’s performance is validated across different industries, with SGM-BRR exhibiting the most optimal performance of RMSE, MAE, and R² in 27, 25, and 27 groups, respectively. Consequently, the model demonstrates broad applicability and stability performance in ESG score prediction.

The mechanism of coal and gas outburst disasters is perplexing, and the evaluation methods of outburst disasters based on various sensitive indicators often have some imprecision and fuzziness. With the concept of accurate and intelligent mining in coal mines proposed in China, selecting quantifiable parameters for machine learning risk prediction can avoid the deviation caused by human subjectivity, and improve the accuracy of coal and gas outburst prediction. Aiming at the shortcomings of the support vector machine (SVM) such as low noise resistance and being prone to be influenced by parameters easily, this research proposed a prediction method based on a grey wolf optimizer to optimize the support vector machine (GWO-SVM). To coordinate the global and local optimization ability of the GWO, Tent Chaotic Mapping and DLH strategies were introduced to improve the optimization ability of the GWO and reduce the local optimal probability. The improved prediction model IGWO-SVM was used to predict the coal and gas outburst. The results showed that this model has faster training speed and higher classification prediction accuracy than the SVM and GWO-SVM models, which the accuracy rate reaching 100%. Finally, to obtain the correlation between the parameters of the coal and gas outburst prediction parameters, the random forest algorithm was used for training, and the three parameters with the highest feature importance were selected to rebuild the data set for machine learning. The accuracy of the IGWO-SVM outburst prediction model based on Random Forest was still 100%. Therefore, even if some prediction parameters are missing, the outburst can still be effectively predicted by using the RF-IGWO-SVM model, which is beneficial for the model application and underground safety management.

This study investigates the suppression effect of ultrafine CaCO₃ powder on gas explosions through a series of experiments conducted in a medium-scale explosion tube (9.6 m in length) with varying concentrations of ultrafine CaCO₃. The gas explosion suppression concentration was established at 9.5%. The results indicate that, at concentrations of 125 g/m³ and 300 g/m³, ultrafine CaCO₃ powder significantly reduced the flame propagation speed of the gas explosion. Among the concentrations that did not fully suppress the gas explosion, 20 g/m³ effectively mitigated the flame propagation speed and overpressure throughout the entire process. The concentration of 75 g/m³ demonstrated a suppression–promotion–suppression–promotion pattern regarding the flame propagation rate, with significant fluctuations observed throughout the process. While the maximum suppression rate reached 60.9%, the maximum promotion rate was 131.12%. At 10 g/m³, the flame propagation rate initially increased before decreasing, with the most effective suppression observed within the first 6.6 m, where flame propagation suppression increased rapidly from 58.13% to 74.82%, and the peak explosion overpressure was reduced by up to 62.94%. These findings contribute valuable insights for the development of effective gas explosion suppression strategies in mining environments.