Conformal Symbolic Regression Applied to Rock Blasting Vibration Prediction

Rock blasting is a crucial technique employed in mining construction and quarrying industries to fragment large rocks into smaller, more manageable pieces. However, it often generates ground vibrations that can pose significant challenges. This study explores the application of Conformal Prediction (CP) in the context of rock blasting to develop predictive models that estimate blasting-induced vibrations and quantify the associated uncertainties. Various machine learning models were integrated within the CP framework using MAPIE and CREPES libraries. The results demonstrated the effectiveness of CP in providing statistically reliable prediction intervals, making it particularly useful for applications where the stakes of underestimating uncertainty can be high.