Application of Bayesian Neural Network (BNN) for the Prediction of Blast-Induced Ground Vibration

Rock blasting is one of the most common and cost-effective excavation techniques. However, rock has various negative environmental effects, such as air overpressure, fly rock, ground vibration. Ground vibration hazardous these inevitable impacts since it a impact not only on environment surrounding area but also human population itself. The PPV critical base parameter practice for understanding, evaluating, predicting in terms velocity. This study aims to predict blast-induced Mikurahana quarry, using Bayesian neural network (BNN) four machine learning techniques, namely, gradient boosting, k-neighbors, decision tree, random forest. proposed models were developed eight input parameters, output, hundred datasets. assessment suitability model comparison others was conducted by different performance evaluation metrics, R, RMSE, MSE. Hence, this compared performances BNN with regression analyses, found that result from superior, lower error: R = 0.94, RMSE 0.17, MSE 0.03. Finally, after models, SHAP performed describe importance models’ features avoid black box issue.