露天台阶爆破块度参数的RBF-SVM预测模型

The RBF-SVM Prediction Model of Bench Blasting Fragmentation Parameters

为了分析德兴铜矿台阶爆破块度参数,利用核函数为Gauss径向基核函数的支持向量回归算法(RBF-SVM)对9组样本的爆后的块度参数进行了模型构造,并对2组测试样本进行了预测。从预测结果可以看出,预测结果的平均相对误差仅为3.03%,说明该模型的可靠性,模型具有很好的预测精度,具有一定的推广价值。

In order to analysis the fragmentation parameters of bench blasting in Dexing copper mine, made the model structure basic on 9 samples blasting parameters by using Gauss radial basis kernel function of support vector machine （RBF-SVM）, and predicted 2 sets of test samples. The prediction results showed the average relative error was 3.03% between the measurements of the test samples and the prediction. And it was proved RBF - SVM model had reliable predicted values and good accuracy measured value, which had an certain value for promotion.