摘要
为了准确预测小样本条件下露天矿山岩石的爆破块度,并得到小样本条件下预测露天矿山爆破块度的有效方法,借助最小二乘支持向量机工具(LS-SVMlab)构建基于最小二乘支持向量机回归(LS-SVR)预测模型并合理优化模型参数。分别使用15组露天矿山爆破数据和35组爆破数据作为小样本容量和正常样本容量,对模型的预测精度进行检验。结果表明:两种样本容量下LS-SVR预测模型的预测结果精度都比同样本容量下人工神经网络(ANN)回归预测的结果精度更高,说明所提出的LS-SVR模型适用于预测露天矿山爆破块度,并且在小样本条件下更具优势。
In order to accurately predict the rock blasting fragmentation of the surface mine with small sample data, an effective method was put forward. The prediction model was built up based on LS-SVM lab and optimized model parameters. 15 groups of surface mine blasting parameters and 35 groups of blasting parameters,as small sam- ple capacity and normal capacity respectively, were separately used to test the prediction accuracy of the model. The results show that LS-SVR has better prediction accuracy than ANN with same sample capacity. The results indicate that the LS-SVR model is suitable for predicting blasting fragmentation of surface mine and prior to regression analysis with small sample data.
出处
《爆破》
CSCD
北大核心
2016年第3期36-40,共5页
Blasting
基金
国家科技支撑计划项目(2013BAB02B05)
关键词
支持向量机
最小二乘支持向量机回归
LS-SVMlab
岩石块度
小样本预测
support vector machines
least squares support vector machines regression
LS-SVMIab
rock flagmentation
prediction with few observations
