回采巷道围岩分类的支持向量机方法

Method of Support Vector Machines for Classifying Surrounding Rocks of Gateways

采用支持向量机“一对一”分类方法,研究了回采巷道围岩分类问题.选择6项影响回采巷道围岩分类的主要指标,利用30组巷道围岩数据作为学习样本,建立了回采巷道围岩分类的支持向量机模型.应用该模型对平顶山矿区4条回采巷道围岩进行了分类预测.结果表明,预测类别与实际类别完全吻合,这说明支持向量机方法可以很好地描述巷道围岩影响指标与围岩类别之间复杂的非线性关系.

The ＂one to one＂ method of support vector machines was used in the research on classifying surrounding rocks. Based on 30 typical samples of gateways including six major factors, a new model of support vector machines for the classification of surrounding rocks of gateways was constructed. The classification of 4 gateways in Pingdingshan mining area was predicted by using this model, indicating that the predicted classification was consistent with the actually measured one. Therefore, the complicated nonlinear relation between classification of surrounding rock gateways and its influencing factors is perfectly presented by the model of support vector machines.