ABSTRACT
This research article explores an innovative method for predicting rockbursts using Machine Learning (ML) algorithms. The study utilizes a decision tree (DT) algorithm and tests two distinct approaches: (1) utilizing a DT model for each rock type (DT-RT) and (2) developing a single DT model (Unique-DT) for all rock types. The dataset comprises 210 records from China, Canada, the United States, Japan, and Italy and includes five input variables for training and testing. The effectiveness of the DT models is compared to other ML algorithms, including Random Forest (RF) and Gradient Boosting (AdaboostM1). The results show that the Unique-DT model performs well and has an F1 score of 0.65 in predicting rockburst conditions. Although RF and AdaboostM1 with F1 (0.66) outperform the Unique-DT model slightly, the Unique-DT model is recommended due to its superior ease of use, effectiveness, and accuracy.
