以矿山台阶爆破块度分布分析为例,建立了爆破块度预测的SVM模型。预测模型与实测结果相比,平均相对误差为6.67%,与BP神经网络模型、R-R分布及G-G-S经验模型相比,SVM模型预测爆破块度具有明显的优越性和可靠性。
With example analysis of size distribution of rock fragment of bench blasting,a support vector machine(SVM) prediction model of rock fragment size for bench blasting was established.The predicted results obtained by the SVM prediction model were in accordance with measured results,the average relative error is 6.67%.Compare with BP neural network model,R-R distribution and G-G-S experience model,the SVM prediction model has obvious superiority and reliability in predicting the rock fragment size for blasting.