Classifying Rockburst Events and Intensity in Underground Mines using Grey Wolf Optimization–Support Vector Machine and Extreme Gradient Boosting

Rockbursts are destructive accidents that often occur in underground mines. With the advancement of technology, machine learning has emerged as an alternative solution that can be used for rockburst mitigation. In this research, we classify rockburst events and their intensities in underground mines using two machine learning models: grey wolf optimization–support vector machine (GWO–SVM) and extreme gradient boosting (XGBoost). Rockburst events are classified into two categories: Existent and None. Meanwhile, the intensities are classified into three categories: weak, moderate, and strong. The implementation used 476 cases of rockbursts with six variables: maximum tangential stress, uniaxial compressive strength, uniaxial tensile strength, stress coefficient, rock brittleness coefficient, and elastic strain index. Both models can better predict the “Existent” rockburst class with a “Weak” intensity compared with the other intensity classes. The performances of the models are evaluated using different proportions of training data, ranging from 50% to 90%. Both models have the same performance for rockburst event classification with 97.53% accuracy, 0.9444 precision, 0.9846 recall, and 0.9628 F1-score. Meanwhile, for intensity classification, XGBoost outperforms GWO-SVM with its 88.24% accuracy, 0.8413 precision, 0.9137 recall, and 0.8651 F1-score.