一种新型的矿井突水分析与预测的支持向量机模型

针对矿井突水样本数少,信息不完整的特点,提出了矿井突水分析的线性核H-SVMs模型。推导模型的理论推广误差公式,设计自顶向下基于SVM最大间隔逐层分类构造H-SVMs的新方法,并应用于实际的矿井突水预测。实验结果表明,线性核H-SVMs模型结构简单、泛化能力强,不仅能很好地预测矿井突水,而且其层次结构能正确反映突水的等级关系,各判别函数的法向量还可以指示各突水影响因素的权重,通过判决函数能有效分析突水影响因素并提取突水预测规则,为矿井突水预测提供了新的方法。

A Novel SVM Model for the Analysis and Prediction of Water Inrush from Coal Mine

In order to analyze such water inrush data with small samples and low accuracy,a linear kernel H-SVMs(Hierarchical Support Vector Machines) model was presented.Firstly,a model was deduced to evaluate the generalization power of H-SVMs,then,a novel method to build H-SVMs was put forward,in which the separation margin calculated by SVM was taken as classification index from top down to bottom and the classes of water inrush samples at each H-SVMs node were dichotomized by the SVM whose separation margin was maximum.Finally,the H-SVMs model was applied to the prediction of mine water inrush.Experimental results show the novel method has a simple structure and a good generalization performance,it can not only predict the scale of water inrush correctly,but also its tree structure can denote the hiberarchy of water inrush.Moreover,the normal vector parameters in the decision functions can describe the weights of the factors related to the mine water inrush and the prediction rules are abstracted from the decision functions by analyzing the decision functions.This study provides a novel scientific method for the prediction of water inrush.