Rock Fragmentation Prediction Using an Artificial Neural Network and Support Vector Regression Hybrid Approach

While empirical rock fragmentation models are easy to parameterize for blast design, they usually prone errors, resulting in less accurate fragment size prediction. Among other shortfalls, these may be unable accurately account the nonlinear relationship that exists between input and output parameters. Machine learning (ML) algorithms potentially able better relationship. To this end, we assess potential of multilayered artificial neural network (ANN) support vector regression (SVR) ML techniques Using geometric, explosives, parameters, build ANN SVR predict mean size. Both yield satisfactory results show higher performance when compared with conventional Kuznetsov model. We further demonstrate an automated means analyzing a varied number hidden layers using Bayesian optimization Keras Python library.