Prediction of rock fragmentation due to blasting operation using Support Vector Machine approac

چکیده مقاله

The main purpose of blasting operations in surface mining is the rock fragmentation, and is considered to be essential to the success of mining operations. To construct the model, parameters such as burden, spacing, hole depth, stemming and powder factor have been considered as input parameters. Support vector machine (SVM) is a novel machine learning technique usually considered as a robust artificial intelligence method in classification and regression tasks. In this paper, rock fragmentation of the Anguran mine has been predicted by developing a model using support vector machine approach. After using the 180 sets of the measured data in Anguran mine for training and testing, the model was applied to 36 data sets as test set for validation of the trained support vector machine model. To investigate the suitability of this approach, predictions by SVM model have been compared with the multivariate regression analysis (MVRA), too. The performance of these models was assessed through the root mean square error, correlation coefficient (R2) and mean absolute percentage error. The results show that R2 between predicted and measured values of the prediction set for SVM model is 0.94 compared to 0.63 of the developed MVRA model. As a result, it was found that the constructed SVM exhibited a higher performance than the empirical method and multivariate regression analysis for rock fragmentation prediction

دریافت لینک دانلود مقاله