A novel ensemble machine learning model to predict mine blasting-induced rock fragmentation

摘要

In production blasting, the primary goal is to produce an appropriate fragmentation, whereas an improper fragmentationis one of the most common side effects induced by these events. This investigation aims at predicting rock fragmentationthrough a new ansemble technique, namely light gradient-boosting machine (LightGBM) with its hyper-parameters thatwere tuned using a powerful optimization algorithm, i.e., the Jellyfish Search Optimizer (JSO). The hybrid JSO-LightGBMis responsible for obtaining the highest possible performance from a combination of these two models where the useddatabase is collected from the Sungun copper mine, Iran. Some blasting pattern parameters such as stemming and spacingwere used as input variables while the mean fragment size (D_(50)), which is a valid indicator for rock fragmentation studies,was considered an output variable. As a result, the coefficient of determination (R~2) of 0.990 on the training set and R~2 of0.996 on the testing set confirmed that the newly developed JSO-LightGBM model has a powerful capability for predictingrock fragmentation, and it can be used as a new methodology in this field. Furthermore, the correlations between the inputvariables and target output by the Shapley Additive exPlanations technique showed that the powder factor has the mostsignificant impact on fragmentation.