Accurate prediction of short-term rockburst has a significant role in improving the safety of workers in mining and geotechnical projects. The rockburst occurrence is nonlinearly correlated with its influencing factors that guarantee imprecise predicting results by employing the traditional methods. In this study, three approaches including including t-distributed stochastic neighbor embedding (t-SNE), K-means clustering, and extreme gradient boosting (XGBoost) were employed to predict the short-term rockburst risk. A total of 93 rockburst patterns with six influential features from micro seismic monitoring events of the Jinping-II hydropower project in China were used to create the database. The original data were randomly split into training and testing sets with a 70/30 splitting ratio. The prediction practice was followed in three steps. Firstly, a state-of-the-art data reduction mechanism t-SNE was employed to reduce the exaggeration of the rockburst database. Secondly, an unsupervised machine learning, i.e., K-means clustering, was adopted to categorize the t-SNE dataset into various clusters. Thirdly, a supervised gradient boosting machine learning method i.e., XGBoost was utilized to predict various levels of short-term rockburst database. The classification accuracy of XGBoost was checked using several performance indices. The results of the proposed model serve as a great benchmark for future short-term rockburst levels prediction with high accuracy.

The uniaxial compressive strength (UCS) of rock is one of the essential data in engineering planning and design. Correctly testing UCS of rock to ensure its accuracy and authenticity is a prerequisite for assuring the design of any rock engineering project. UCS of rock has a broad range of applications in mining, geotechnical, petroleum, geomechanics, and other fields of engineering. The application of the gradient boosting machine learning algorithms has been rarely used, especially for UCS prediction, and has performed well, based on the relevant literature of the study. In this study, four gradient boosting machine learning algorithms, namely, gradient boosted regression (GBR), Catboost, light gradient boosting machine (LightGBM), and extreme gradient boosting (XGBoost), were developed to predict the UCS in MPa of soft sedimentary rocks of the Block‐IX at Thar Coalfield, Pakistan, using four input variables such as wet density (ρ_w) in g/cm³; moisture in %; dry density (ρ_d) in g/cm³; and Brazilian tensile strength (BTS) in MPa. Then, 106‐point dataset was allocated identically for each algorithm into 70% for the training phase and 30% for the testing phase. According to the results, the XGBoost algorithm outperformed the GBR, Catboost, and LightGBM with coefficient of correlation (R²) = 0.99, mean absolute error (MAE) = 0.00062, mean square error (MSE) = 0.0000006, and root mean square error (RMSE) = 0.00079 in the training phase and R² = 0.99, MAE = 0.00054, MSE = 0.0000005, and RMSE = 0.00069 in the testing phase. The sensitivity analysis showed that BTS and ρ_w are positively correlated, and the moisture and ρ_d are negatively correlated with the UCS. Therefore, in this study, the XGBoost algorithm was shown to be the most accurate algorithm among all the investigated four algorithms for UCS prediction of soft sedimentary rocks of the Block‐IX at Thar Coalfield, Pakistan.

Drilling rate index (DRI) is a fundamental parameter in the investigation of rock drillability, as drillability is considered one of the main problems in rock engineering. Several researchers have continuously tried to analyze and correlate rock DRI, but the problem remains unchanged. This study elucidates the machine learning approaches, namely long short term memory (LSTM), simple recurrent neural network (RNN) and random forest algorithm (RFA) to predict DRI of rocks using multivariate inputs, that is, uniaxial compressive strength in MPa; Brazilian tensile strength (BTS) in MPa; brittleness value (S₂₀); Sievers’ J value (S_j); modulus ratio (MR); shore hardness (SH), porosity (n) in %; shimazeks F abrasitivity in N/mm; and equivalent quartz content in %. For all proposed methods, the original dataset was divided into 70% for training and the remaining 30% for testing. Next, the performance indices, such as correlation coefficient (R²), root mean square error (RMSE), variance accounts for (VAF) and a-20 index of each proposed method were determined to examine the accuracy of the predicted data. In this study, according to the results of LSTM, simple RNN and RFA methods, the LSTM revealed the best prediction output for DRI with the strongest R², the lowest RMSE, the largest VAF and an appropriate a-20 index values as 0.999, 0.13416, 0.997, and 0.999 in the training stage and 0.998, 0.19479, 0.996, and 0.997 in the testing stage, respectively. Therefore, LSTM is an applicable machine learning approach that can be applied to accurately predict the DRI.

The rockburst phenomenon is the major source of the high number of casualties and fatalities during the construction of deep underground projects. Rockburst poses a severe hazard to the safety of employees and equipment in subsurface mining operations. It is a hot topic in recent years to examine and overcome rockburst risks for the safe installation of deep urban engineering designs. Therefore, for a cost-effective and safe underground environment, it is crucial to determine and predict rockburst intensity prior to its occurrence. A novel model is presented in this study that combines unsupervised and supervised machine learning approaches in order to predict rockburst risk. The database for this study was built using authentic microseismic monitoring occurrences from the Jinping-II hydropower project in China, which consists of 93 short-term rockburst occurrences with six influential features. The prediction process was succeeded in three steps. Firstly, the original rockburst database's magnification was reduced using a state-of-the-art method called isometric mapping (ISOMAP) algorithm. Secondly, the dataset acquired from ISOMAP was categorized using the fuzzy c-means algorithm (FCM) to reduce the minor spectral heterogeneity impact in homogenous areas. Thirdly, K-Nearest neighbor (KNN) was employed to anticipate different levels of short-term rockburst datasets. The KNN's classification performance was examined using several performance metrics. The proposed model correctly classified about 96% of the rockbursts events in the testing datasets. Hence, the suggested model is a realistic and effective tool for evaluating rockburst intensity. Therefore, the proposed model can be employed to forecast the rockburst risk in the early stages of underground projects that will help to minimize casualties from rockburst.