Coal and gas outburst prediction data are characterized by high dimensionality and small sample sizes, posing significant challenges to predictive modeling. To address this issue, this study constructed a database of 60 samples comprising seven indicators, including gas pressure, gas content, and coal failure type. The permutation importance method was used for feature dimensionality reduction, selecting five key features (initial velocity of gas emission, coal seam thickness, gas content, gas pressure and coal sturdiness coefficient) to mitigate the impact of weakly correlated features on prediction modeling. A Stacking ensemble model was developed using support vector machine (SVM), random forest (RF), K-nearest neighbor (KNN), logistic regression (LR) and extreme gradient boosting (XGBoost) as base learners and XGBoost as the meta-learner. Bayesian Optimization (BO) was applied for global hyperparameter tuning, resulting in a BO-Stacking ensemble model for coal and gas outburst prediction. The shapely additive explanations (SHAP) method was employed for interpretability analysis of the model's predictions. The results show that the BO-Stacking model, after feature reduction, achieved an accuracy of 92.4 %, an F₁ score of 0.956, a Kappa coefficient of 0.927, and an AUC value of 0.969, outperforming all individual models. The ranking of feature importance was initial velocity of gas emission > gas content > gas pressure > coal sturdiness coefficient > coal seam thickness. The BO-Stacking ensemble learning model demonstrates strong predictive performance and stability, providing a novel approach for coal and gas outburst prediction.

As mining operations delve deeper and mechanization and intelligence levels improve, coal mine disasters are becoming increasingly severe. Consequently, developing effective technology and equipment is crucial to ensure the safety of mining enterprises. This perspective summarizes the technical methods for preventing coal and rock dynamic disasters and controlling dust in coal mines. Furthermore, it provides insights into the future directions of mining disaster prevention techniques and equipment in this field. The aim of this paper is to offer effective disaster prevention strategies, enhance the efficiency and effectiveness of disaster control, and further safeguard the health and safety of miners.

Rock bursts pose a significant risk to coal mine operation safety. Thus, accurately discriminating coal bursting liabilities is crucial for predicting and preventing rock burst events. To better understand the effects of a varying bedding angle on the crack propagation rule, failure mode and bursting liability level of coal and coal-rock combinations, we propose an optimized machine learning-based model. Additionally, uniaxial compressive tests are conducted using PFC3D software on samples with different bedding angles. The results indicate that, among the nine light gradient boosting machine discriminant models constructed using three data preprocessing methods and three parameter optimization algorithms, the optimal model is identified as the particle swarm optimization-light gradient boosting machine discriminant model based on Z-score standardization method, which exhibits the best stability and has a F1-score of 93.6%. Bedding has a significant impact on the failure modes of two kinds of samples, resulting in an evident bedding effect on their bursting liability. The uniaxial compression strength and bursting energy index of both samples show a reduction-rising trend with an increasing bedding dip angle. However, the bursting liability level of these samples is not affected by 0° or 90° bedding dip angle. Therefore, when assessing the bursting liability of samples, the influence of coal seam bedding and its dip angles should be thoroughly considered.