Abstract:Blasting in bench operations is widely used in mining and infrastructure construction, but its vibration effects pose a serious threat to the safety of adjacent facilities. Accurate prediction of the peak vibration velocity induced by blasting is crucial for ensuring engineering safety. Addressing the limited generalization ability of traditional empirical formulas and single machine learning models, this study proposes an integrated prediction model that combines the Sand Cat Swarm Optimization (SCSO) algorithm with Categorical Boosting (CatBoost) to enhance prediction accuracy. Experimental results demonstrate that optimizing the CatBoost model with the SCSO algorithm significantly improves prediction accuracy. This model performs excellently across multiple datasets, providing a new and effective method for predicting peak vibration velocities in bench blasting and having potential for application in other related projects.
